GPU for the Web Working Group - Activity in 2022
No matches.
Lifecycle
-
1 charter started.
-
Chartered until 30 November 2024.
-
TR publications
-
8 documents published.
-
WebGPU published 5 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 3 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/gpuweb
-
33 issues opened.
-
[editorial] GPUPipelineError constructor argument `message` should be optional gpuweb/gpuweb#3709, by @kainino0x.
-
Improve syntax authoring experience gpuweb/gpuweb#3708, by @mehmetoguzderin.
-
Increase distinguishing factor between token types gpuweb/gpuweb#3707, by @mehmetoguzderin.
-
Complete the #image-input section of the explainer gpuweb/gpuweb#3703, by @Kangz.
-
Render pipelines with no attachments shouldn't be allowed gpuweb/gpuweb#3702, by @Kangz.
-
Make GPURenderPassDecriptor.colorAttachments optional gpuweb/gpuweb#3701, by @Kangz.
-
`maxBufferSize` questions gpuweb/gpuweb#3700, by @teoxoy.
-
Is it intended that keywords can not be identifiers? gpuweb/gpuweb#3692, by @greggman.
-
Shader compilation is fragile because FXC gpuweb/gpuweb#3691, by @raphlinus.
-
wgsl: a triggering diagnostic can have additional notes expressed as "info" diagnostics. gpuweb/gpuweb#3690, by @dneto0.
-
Explainer is marked as a "living standard" gpuweb/gpuweb#3687, by @Kangz.
-
Increase visibility of technical words by opting into existing synonyms gpuweb/gpuweb#3686, by @mehmetoguzderin.
-
Bump macOS version to 10.15 gpuweb/gpuweb#3685, by @teoxoy.
-
Missing `filterable` texture format definiton gpuweb/gpuweb#3684, by @teoxoy.
-
Pipe WGSL compilation failures in createFooPipeline. gpuweb/gpuweb#3682, by @Kangz.
-
Cannot cast vec2<u32> to vec2<f32> gpuweb/gpuweb#3681, by @PENGUINLIONG.
-
Distinguish between pointer and pointer contents for uniformity requirements on parameters gpuweb/gpuweb#3677, by @jrprice.
-
the native API for webgl2 is opengl es3 or d3d11？ gpuweb/gpuweb#3676, by @yangfangfang1204.
-
Add short-named aliases for matrices (e.g. `mat2x3f`, `mat4x2h`) gpuweb/gpuweb#3673, by @ben-clayton.
-
Restrict grammar of `texture_and_sampler_types` gpuweb/gpuweb#3672, by @teoxoy.
-
Behavior of `GPUSamplerDescriptor.maxAnisotropy` with a non power of two gpuweb/gpuweb#3671, by @teoxoy.
-
Uniform control flow vs textures for flat interpolated varyings gpuweb/gpuweb#3668, by @unconed.
-
beginComputePass/beginRenderPass errors do not invalidate parent encoder gpuweb/gpuweb#3663, by @kainino0x.
-
TR publication broken due to issue with document license/charter gpuweb/gpuweb#3662, by @kainino0x.
-
mapAsync has an incorrect assertion gpuweb/gpuweb#3660, by @kainino0x.
-
wgsl: add sign builtin function on integer scalar and vector gpuweb/gpuweb#3658, by @dneto0.
-
Does it support glPolygonMode and glPolygonOffset? gpuweb/gpuweb#3657, by @pixcai.
-
How can I get the minimum float value? gpuweb/gpuweb#3656, by @javirk.
-
"validating inter-stage interfaces" doesn't handle vertex-only pipelines gpuweb/gpuweb#3653, by @kainino0x.
-
Metal can only create 32 MTLCounterSampleBuffers at once? gpuweb/gpuweb#3652, by @kainino0x.
-
-
19 issues closed.
-
Cannot cast vec2<u32> to vec2<f32> gpuweb/gpuweb#3681.
-
Add short-named aliases for matrices (e.g. `mat2x3f`, `mat4x2h`) gpuweb/gpuweb#3673.
-
Restrict grammar of `texture_and_sampler_types` gpuweb/gpuweb#3672.
-
beginComputePass/beginRenderPass errors do not invalidate parent encoder gpuweb/gpuweb#3663.
-
the native API for webgl2 is opengl es3 or d3d11？ gpuweb/gpuweb#3676.
-
[webgpu] Broken references in WebGPU gpuweb/gpuweb#3617.
-
Confusing note in builtin `length` section gpuweb/gpuweb#3621.
-
Behavior of `GPUSamplerDescriptor.maxAnisotropy` with a non power of two gpuweb/gpuweb#3671.
-
Metal can only create 32 MTLCounterSampleBuffers at once? gpuweb/gpuweb#3652.
-
Primitive clipping at the rasterization stage is inconsistent between APIs gpuweb/gpuweb#3638.
-
TR publication broken due to issue with document license/charter gpuweb/gpuweb#3662.
-
wgsl: add sign builtin function on integer scalar and vector gpuweb/gpuweb#3658.
-
mapAsync has an incorrect assertion gpuweb/gpuweb#3660.
-
Workgroup Broadcast gpuweb/gpuweb#2586.
-
How can I get the minimum float value? gpuweb/gpuweb#3656.
-
Metal seem to support only up to 4096 queries in sample buffers. gpuweb/gpuweb#3648.
-
1 issue reopened.
-
Spec the texture-compression-etc and texture-compression-astc extensions gpuweb/gpuweb#2083, by @lokokung.
-
-
45 issues received 113 new comments.
-
Uniform control flow vs textures for flat interpolated varyings gpuweb/gpuweb#3668 (9 comments by @dneto0, @fintelia, @kdashg, @unconed)
-
Permissions Policy for WebGPU and for unmaskHints gpuweb/gpuweb#3483 (6 comments by @domenic, @ghost, @kainino0x, @Kangz)
-
Behavior of `GPUSamplerDescriptor.maxAnisotropy` with a non power of two gpuweb/gpuweb#3671 (6 comments by @kainino0x, @Kangz, @teoxoy)
-
beginComputePass/beginRenderPass errors do not invalidate parent encoder gpuweb/gpuweb#3663 (5 comments by @austinEng, @kainino0x, @toji)
-
Improve color and display-size readability of diagrams and code blocks and other elements in WGSL spec gpuweb/gpuweb#3649 (4 comments by @kainino0x, @mehmetoguzderin)
-
wgsl: refine range diagnostic filters: where they can be applied, and associated affected ranges gpuweb/gpuweb#3689 (4 comments by @ben-clayton, @dneto0)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (4 comments by @kainino0x, @xfq)
-
Spec says “should warn” for unicode homographs. Should we allow control of that one via diagnostics? gpuweb/gpuweb#3688 (3 comments by @kdashg, @mehmetoguzderin)
-
Restrict grammar of `texture_and_sampler_types` gpuweb/gpuweb#3672 (3 comments by @dneto0, @kdashg, @teoxoy)
-
WGSL reader failure: :7:14 error: no matching call to textureSampleLevel(texture_external, sampler, vec2<f32>) gpuweb/gpuweb#3680 (3 comments by @ben-clayton)
-
TR publication broken due to issue with document license/charter gpuweb/gpuweb#3662 (3 comments by @kainino0x, @tidoust)
-
wgsl: add sign builtin function on integer scalar and vector gpuweb/gpuweb#3658 (3 comments by @dneto0, @kdashg)
-
rg11b10ufloat-renderable should probably allow use as a resolve target gpuweb/gpuweb#3566 (2 comments by @kainino0x, @teoxoy)
-
wgsl: language evolution: managing rollout of core (essentially sugar) language features gpuweb/gpuweb#3149 (2 comments by @kdashg)
-
the native API for webgl2 is opengl es3 or d3d11？ gpuweb/gpuweb#3676 (2 comments by @kainino0x, @munrocket)
-
wgsl: full reference to module-scope variable is uniform, the value stored there may not be gpuweb/gpuweb#3602 (2 comments by @dneto0, @kdashg)
-
Does it support glPolygonMode and glPolygonOffset? gpuweb/gpuweb#3657 (2 comments by @greggman, @Kangz)
-
Render pipelines with no attachments shouldn't be allowed gpuweb/gpuweb#3702 (1 comment by @kainino0x)
-
Add clarification note around variables in the `body` and not visible in the `update_part` of a `for` gpuweb/gpuweb#3187 (1 comment by @dneto0)
-
Distinguish between pointer and pointer contents for uniformity requirements on parameters gpuweb/gpuweb#3677 (1 comment by @kdashg)
-
Primitive clipping at the rasterization stage is inconsistent between APIs gpuweb/gpuweb#3638 (1 comment by @Kangz)
-
createPipelineAsync() doesn't reject its promise, even if the created pipeline is invalid gpuweb/gpuweb#3296 (1 comment by @kainino0x)
-
Spec the texture-compression-etc and texture-compression-astc extensions gpuweb/gpuweb#2083 (1 comment by @lokokung)
-
Almost no PowerVR devices satisfy minimum Vulkan requirements gpuweb/gpuweb#3631 (1 comment by @kainino0x)
-
Optional feature request: filterable 32-bit float textures gpuweb/gpuweb#3601 (1 comment by @kainino0x)
-
-
22 pull requests opened.
-
Make GPUPipelineError constructor `message` pseudo-optional gpuweb/gpuweb#3710, by @kainino0x.
-
Fix include gpuweb/gpuweb#3706, by @mehmetoguzderin.
-
Make the explainer a "Living Document" gpuweb/gpuweb#3705, by @kainino0x.
-
Remove spurious note on depthStoreOp gpuweb/gpuweb#3704, by @kainino0x.
-
Complete several algorithms that enqueue commands gpuweb/gpuweb#3699, by @toji.
-
Reset [[pending_map]] to null in the RangeError case gpuweb/gpuweb#3698, by @kainino0x.
-
Fill out the compilationInfo() algorithm gpuweb/gpuweb#3695, by @toji.
-
Fix typo for vec2<u32> to vec2<f32> conversion gpuweb/gpuweb#3693, by @dneto0.
-
Add diagnostics, derivative uniformity problems trigger diagnostics gpuweb/gpuweb#3683, by @dneto0.
-
Small typo in explainer gpuweb/gpuweb#3679, by @eliemichel.
-
Update WGSL links gpuweb/gpuweb#3675, by @toji.
-
wgsl: add predeclared type aliases for matrices gpuweb/gpuweb#3674, by @dneto0.
-
[editorial] wgsl builtin function notes gpuweb/gpuweb#3670, by @munrocket.
-
Added a notes about language/direction metadata gpuweb/gpuweb#3669, by @toji.
-
Allow bgra8unorm in pipeline layout be compatible with rgba8unorm in … gpuweb/gpuweb#3667, by @Jiawei-Shao.
-
text/wgsl media type registration is a normative part of the WGSL spec. gpuweb/gpuweb#3666, by @dneto0.
-
begin pass: only generate error on incorrect state gpuweb/gpuweb#3665, by @kainino0x.
-
Clarify powerPreference may also affect the power configuration on se… gpuweb/gpuweb#3664, by @Jiawei-Shao.
-
Fix mapping state assertions in mapAsync gpuweb/gpuweb#3661, by @kainino0x.
-
Add integer overloads for sign builtin function gpuweb/gpuweb#3659, by @dneto0.
-
Link to WGSL definition of line breaks gpuweb/gpuweb#3654, by @toji.
-
Check user IO equality only if the fragment stage is present. gpuweb/gpuweb#3651, by @Kangz.
-
-
22 pull requests merged.
-
Use task sources for device events and canvas texture expiry gpuweb/gpuweb#3632.
-
Reject create*PipelineAsync on failure gpuweb/gpuweb#3523.
-
Fix include gpuweb/gpuweb#3706.
-
WGSL syntax styling upgrade with initial notation section gpuweb/gpuweb#3543.
-
Reset [[pending_map]] to null in the RangeError case gpuweb/gpuweb#3698.
-
Fill out the compilationInfo() algorithm gpuweb/gpuweb#3695.
-
Fix typo for vec2<u32> to vec2<f32> conversion gpuweb/gpuweb#3693.
-
wgsl: add predeclared type aliases for matrices gpuweb/gpuweb#3674.
-
begin pass: only generate error on incorrect state gpuweb/gpuweb#3665.
-
Small typo in explainer gpuweb/gpuweb#3679.
-
Update WGSL links gpuweb/gpuweb#3675.
-
Clarify powerPreference may also affect the power configuration on se… gpuweb/gpuweb#3664.
-
Added a notes about language/direction metadata gpuweb/gpuweb#3669.
-
[editorial] Remove note about non-meaningful result of scalar length gpuweb/gpuweb#3622.
-
Add integer overloads for sign builtin function gpuweb/gpuweb#3659.
-
text/wgsl media type registration is a normative part of the WGSL spec. gpuweb/gpuweb#3666.
-
Fix mapping state assertions in mapAsync gpuweb/gpuweb#3661.
-
Add workgroupuniform load built-in and analysis gpuweb/gpuweb#3586.
-
Link to WGSL definition of line breaks gpuweb/gpuweb#3654.
-
Check user IO equality only if the fragment stage is present. gpuweb/gpuweb#3651.
-
Fix issues in pass end() validation gpuweb/gpuweb#3626.
-
Reduce query set count limit to 4096 gpuweb/gpuweb#3650.
-
-
4 pull requests closed.
-
Allow bgra8unorm in pipeline layout be compatible with rgba8unorm in … gpuweb/gpuweb#3667.
-
Add uniformity analysis opt-out for derivative-using builtins gpuweb/gpuweb#3644.
-
Expire the current texture with a microtask gpuweb/gpuweb#3337.
-
wgsl: clarify function domains gpuweb/gpuweb#3646.
-
-
39 pull requests received 153 new comments.
-
Add diagnostics, derivative uniformity problems trigger diagnostics gpuweb/gpuweb#3683 (41 comments by @alan-baker, @dneto0, @jimblandy, @jrprice, @kdashg, @mehmetoguzderin)
-
WGSL: allow bgra8unorm as texel format gpuweb/gpuweb#3640 (30 comments by @dneto0, @jenatali, @Jiawei-Shao, @kainino0x, @Kangz, @kdashg, @litherum, @mehmetoguzderin)
-
Complete several algorithms that enqueue commands gpuweb/gpuweb#3699 (18 comments by @kainino0x, @toji)
-
Added a notes about language/direction metadata gpuweb/gpuweb#3669 (7 comments by @kainino0x, @toji)
-
Add workgroupuniform load built-in and analysis gpuweb/gpuweb#3586 (7 comments by @alan-baker, @dneto0)
-
wgsl: add predeclared type aliases for matrices gpuweb/gpuweb#3674 (6 comments by @dneto0, @jrprice, @kdashg)
-
WGSL syntax styling upgrade with initial notation section gpuweb/gpuweb#3543 (4 comments by @mehmetoguzderin)
-
begin pass: only generate error on incorrect state gpuweb/gpuweb#3665 (4 comments by @kainino0x, @toji)
-
Add uniformity analysis opt-out for derivative-using builtins gpuweb/gpuweb#3644 (3 comments by @dneto0, @kdashg)
-
Require render and resolve to have same base format gpuweb/gpuweb#3363 (3 comments by @kainino0x, @lokokung)
-
Use task sources for device events and canvas texture expiry gpuweb/gpuweb#3632 (1 comment by @kainino0x)
-
Allow bgra8unorm in pipeline layout be compatible with rgba8unorm in … gpuweb/gpuweb#3667 (1 comment by @kainino0x)
-
text/wgsl media type registration is a normative part of the WGSL spec. gpuweb/gpuweb#3666 (1 comment by @dneto0)
-
Extend lifetime of GPUExternalTexture imported from video element gpuweb/gpuweb#2302 (1 comment by @lokokung)
-
workgroupStorageUsed is sum of values, each rounded up to multiple of 16 gpuweb/gpuweb#3522 (1 comment by @lokokung)
-
Only allow defined GPUShaderStage bits in bind group layout entries gpuweb/gpuweb#3500 (1 comment by @lokokung)
-
Validation errors on out-of-bounds pipeline constants, and clearValues gpuweb/gpuweb#3481 (1 comment by @lokokung)
-
Check user IO equality only if the fragment stage is present. gpuweb/gpuweb#3651 (1 comment by @kainino0x)
-
Revert "Make getBGL return a "copy" instead of a "reference"" gpuweb/gpuweb#3455 (1 comment by @lokokung)
-
Matching rules for fragment shader output and color attachments gpuweb/gpuweb#3408 (1 comment by @lokokung)
-
gpuweb/cts
-
18 issues opened.
-
Show some progress when running the CTS gpuweb/cts#2099, by @greggman.
-
Support Dark Mode gpuweb/cts#2097, by @greggman.
-
Fix wrong _errorLine in operation/shader_module/compilation_info.spec.ts gpuweb/cts#2087, by @Jiawei-Shao.
-
allow specifying a default powerPreference gpuweb/cts#2083, by @greggman.
-
Use textureContentIsOKByT2B in copyTextureToTexture expectations gpuweb/cts#2081, by @austinEng.
-
webgpu:web_platform,copyToTexture,* lacks of coverage on subrect copy gpuweb/cts#2079, by @shaoboyan.
-
wgsl: add short predeclared type aliases for matrix types. gpuweb/cts#2074, by @dneto0.
-
op: test comparison sampling results and filtering gpuweb/cts#2072, by @kainino0x.
-
op: sampling,lod_clamp gpuweb/cts#2071, by @kainino0x.
-
op: sampling,filter_mode gpuweb/cts#2070, by @kainino0x.
-
val: *_pipeline,override,conversions gpuweb/cts#2060, by @shrekshao.
-
wgsl: add signed integer overloads for the sign builtin gpuweb/cts#2057, by @dneto0.
-
val:render_pass,render_pass_descriptor:resolveTarget,different_format gpuweb/cts#2049, by @kainino0x.
-
op: create large numbers of different types of objects gpuweb/cts#2048, by @kainino0x.
-
Update expectation values in webgpu:web_platform,external_texture,video:* gpuweb/cts#2042, by @shaoboyan.
-
-
11 issues closed.
-
Check the validation of getBindGroupLayout method gpuweb/cts#1891.
-
val: Handle missing validations in copyT2B and copyB2T gpuweb/cts#1799.
-
op: Add more operation test for index format behavior. gpuweb/cts#1968.
-
op: render_pass,stencil_operation gpuweb/cts#1977.
-
Remove infinity cases from numeric testing code gpuweb/cts#2020.
-
val: mapAsync,state,destroyed gpuweb/cts#929.
-
[wgsl] padding bytes must not be accessed gpuweb/cts#1498.
-
-
2 issues reopened.
-
op: render_pass,stencil_operation gpuweb/cts#1977, by @Gyuyoung.
-
Debug marker validation and operation tests gpuweb/cts#1928, by @Gyuyoung.
-
-
16 issues received 32 new comments.
-
Use textureContentIsOKByT2B in copyTextureToTexture expectations gpuweb/cts#2081 (3 comments by @austinEng, @kainino0x)
-
Update expectation values in webgpu:web_platform,external_texture,video:* gpuweb/cts#2042 (2 comments by @bjjones, @shaoboyan)
-
val: setBindGroup required for a bind group layout that exists but is empty gpuweb/cts#1903 (1 comment by @Gyuyoung)
-
43 pull requests opened.
-
Put subcase in the stack instead of the message gpuweb/cts#2103, by @kainino0x.
-
Fix silencing of DOMExceptions gpuweb/cts#2102, by @kainino0x.
-
Improve Buffer.mapAsync tests gpuweb/cts#2101, by @takahirox.
-
Add some progress gpuweb/cts#2100, by @greggman.
-
Support dark mode gpuweb/cts#2098, by @greggman.
-
Test all stencil formats, and address some nits in depth_bias/stencil tests gpuweb/cts#2096, by @kainino0x.
-
op: Implement 'depth_bias_24bit_format' test in depth_bias.spec.ts gpuweb/cts#2095, by @Gyuyoung.
-
op: Update 'stencil_compare_func' test in stencil.spec.ts gpuweb/cts#2094, by @Gyuyoung.
-
op: Implement 'stencil_depthFailOp_operation' test in stencil.spec.ts gpuweb/cts#2092, by @Gyuyoung.
-
op: Update 'stencil_test_fail' test gpuweb/cts#2091, by @Gyuyoung.
-
op: Implement 'filter_mode,minFilter' test in filter_mode.spec.ts gpuweb/cts#2090, by @Gyuyoung.
-
Fix wrong _errorLine in operation/shader_module/compilation_info.spec.ts gpuweb/cts#2088, by @Jiawei-Shao.
-
Prefix subcase recorder error stacks with the subcase name gpuweb/cts#2086, by @austinEng.
-
fix requestAdapter test. gpuweb/cts#2085, by @greggman.
-
Add a powerPreference option gpuweb/cts#2084, by @greggman.
-
Add 'skipUndefined' Comparator gpuweb/cts#2082, by @zoddicus.
-
Buffer.mapAsync pending map error shouldn't be a validation error gpuweb/cts#2080, by @takahirox.
-
Remove usage of Case Cache in `pack2x16float` gpuweb/cts#2078, by @zoddicus.
-
Expect DataCloneError when transferring a detached mapped ArrayBuffer gpuweb/cts#2076, by @austinEng.
-
op: Introduce depth_bias.spec.ts file with a 'depth_bias' test gpuweb/cts#2075, by @Gyuyoung.
-
Adds explicit check that mapAsync rejects when buffer is destroyed. gpuweb/cts#2069, by @lokokung.
-
Add code coverage collection feature to the server gpuweb/cts#2068, by @ben-clayton.
-
op: Implement 'depth_test_fail' test in depth.spec.ts gpuweb/cts#2067, by @Gyuyoung.
-
Filter oob const gpuweb/cts#2066, by @zoddicus.
-
Also test depthWriteEnabled:true in depth_write_disabled gpuweb/cts#2065, by @kainino0x.
-
op: Implement 'depth_disabled' test in depth.spec.ts gpuweb/cts#2064, by @Gyuyoung.
-
op: Implement 'stencil_reference_initialized' test in stencil.spec.ts gpuweb/cts#2063, by @Gyuyoung.
-
Implement generate*Cases functions and use them gpuweb/cts#2062, by @zoddicus.
-
Add val:*_pipeline,overrides,value,type_error/validation_error gpuweb/cts#2061, by @shrekshao.
-
Remove attempting to emit infinities in WGSL gpuweb/cts#2059, by @zoddicus.
-
Remove infinities from all f32 test inputs gpuweb/cts#2056, by @zoddicus.
-
op: Implmenet 'stencil_pass_operation' test in stencil.spec.ts gpuweb/cts#2055, by @Gyuyoung.
-
Replace usage of vector inputs with fullF32Range for arithmetic tests gpuweb/cts#2053, by @zoddicus.
-
Add earlyRejection parameter to testMapAsyncCall() gpuweb/cts#2052, by @takahirox.
-
wgsl: Add tests for preservation of padding bytes gpuweb/cts#2051, by @jrprice.
-
op: Implement 'depth_disabled' test in depth.spec.ts gpuweb/cts#2050, by @Gyuyoung.
-
op: Implement 'stencil_basic' test gpuweb/cts#2047, by @Gyuyoung.
-
WIP: Remove infinities from values for f32 tests gpuweb/cts#2045, by @zoddicus.
-
op: Implement stencil_fail_operation test in stencil.spec.ts gpuweb/cts#2044, by @Gyuyoung.
-
op: Implement stencil_read_write_mask test in stencil.spec.ts gpuweb/cts#2043, by @Gyuyoung.
-
-
37 pull requests merged.
-
Add some progress gpuweb/cts#2100.
-
Support dark mode gpuweb/cts#2098.
-
Buffer.mapAsync pending map error shouldn't be a validation error gpuweb/cts#2080.
-
op: Update 'stencil_compare_func' test in stencil.spec.ts gpuweb/cts#2094.
-
op: Implement 'depth_bias_24bit_format' test in depth_bias.spec.ts gpuweb/cts#2095.
-
Add earlyRejection parameter to testMapAsyncCall() gpuweb/cts#2052.
-
op: Update 'stencil_test_fail' test gpuweb/cts#2091.
-
op: Introduce depth_bias.spec.ts file with a 'depth_bias' test gpuweb/cts#2075.
-
Prefix subcase recorder error stacks with the subcase name gpuweb/cts#2086.
-
fix requestAdapter test. gpuweb/cts#2085.
-
Implement CTS requestAdapter test gpuweb/cts#1058.
-
Add 'skipUndefined' Comparator gpuweb/cts#2082.
-
Remove usage of Case Cache in `pack2x16float` gpuweb/cts#2078.
-
Filter oob const gpuweb/cts#2066.
-
Remove attempting to emit infinities in WGSL gpuweb/cts#2059.
-
Adds explicit check that mapAsync rejects when buffer is destroyed. gpuweb/cts#2069.
-
Add code coverage collection feature to the server gpuweb/cts#2068.
-
Implement generate*Cases functions and use them gpuweb/cts#2062.
-
op: Implement 'depth_test_fail' test in depth.spec.ts gpuweb/cts#2067.
-
Add val:*_pipeline,overrides,value,type_error/validation_error gpuweb/cts#2061.
-
Also test depthWriteEnabled:true in depth_write_disabled gpuweb/cts#2065.
-
op: Implement 'depth_disabled' test in depth.spec.ts gpuweb/cts#2050.
-
op: Implement 'depth_write_disabled' test in depth.spec.ts gpuweb/cts#2024.
-
Remove infinities from all f32 test inputs gpuweb/cts#2056.
-
wgsl: Add tests for preservation of padding bytes gpuweb/cts#2051.
-
op: Implmenet 'stencil_pass_operation' test in stencil.spec.ts gpuweb/cts#2055.
-
op: Implement stencil_read_write_mask test in stencil.spec.ts gpuweb/cts#2043.
-
op: Implement stencil_fail_operation test in stencil.spec.ts gpuweb/cts#2044.
-
update deps for various packages gpuweb/cts#2041.
-
op: Implement 'stencil_compare_func' test gpuweb/cts#2032.
-
-
4 pull requests closed.
-
WIP: Remove infinities from values for f32 tests gpuweb/cts#2045.
-
op: Implement 'depth_disabled' test in depth.spec.ts gpuweb/cts#2064.
-
op: Implement 'stencil_basic' test gpuweb/cts#2047.
-
35 pull requests received 176 new comments.
-
expect command encoder validation error to be at beginRenderPass/beginComputePass time gpuweb/cts#2046 (17 comments by @kainino0x, @shrekshao)
-
Add val:*_pipeline,overrides,value,type_error/validation_error gpuweb/cts#2061 (16 comments by @kainino0x, @shrekshao)
-
op: Introduce depth_bias.spec.ts file with a 'depth_bias' test gpuweb/cts#2075 (11 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'stencil_reference_initialized' test in stencil.spec.ts gpuweb/cts#2063 (10 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'depth_write_disabled' test in depth.spec.ts gpuweb/cts#2024 (8 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'depth_disabled' test in depth.spec.ts gpuweb/cts#2050 (8 comments by @Gyuyoung, @kainino0x)
-
Add code coverage collection feature to the server gpuweb/cts#2068 (7 comments by @ben-clayton, @Kangz)
-
op: Update 'stencil_compare_func' test in stencil.spec.ts gpuweb/cts#2094 (6 comments by @Gyuyoung, @kainino0x)
-
op: Implement stencil_read_write_mask test in stencil.spec.ts gpuweb/cts#2043 (6 comments by @Gyuyoung, @kainino0x)
-
Test beginComputePass/RenderPass after encoder.finish() in call_after_successful_finish gpuweb/cts#2093 (5 comments by @kainino0x)
-
op: Implement 'depth_bias_24bit_format' test in depth_bias.spec.ts gpuweb/cts#2095 (5 comments by @kainino0x)
-
WIP: Remove infinities from values for f32 tests gpuweb/cts#2045 (5 comments by @ben-clayton, @zoddicus)
-
op: Implement stencil_fail_operation test in stencil.spec.ts gpuweb/cts#2044 (5 comments by @Gyuyoung, @kainino0x)
-
Update copyTextureToTexture tests to allow both encodings of -1 for 8-byte snorm formats gpuweb/cts#2073 (4 comments by @austinEng, @kainino0x, @Kangz)
-
Implement generate*Cases functions and use them gpuweb/cts#2062 (4 comments by @ben-clayton, @zoddicus)
-
Buffer.mapAsync pending map error shouldn't be a validation error gpuweb/cts#2080 (3 comments by @kainino0x, @takahirox)
-
op: Implement 'filter_mode,minFilter' test in filter_mode.spec.ts gpuweb/cts#2090 (2 comments by @Gyuyoung, @kainino0x)
-
Replace usage of vector inputs with fullF32Range for arithmetic tests gpuweb/cts#2053 (2 comments by @ben-clayton, @zoddicus)
-
op: Implmenet 'stencil_pass_operation' test in stencil.spec.ts gpuweb/cts#2055 (2 comments by @Gyuyoung, @kainino0x)
-
Test all stencil formats, and address some nits in depth_bias/stencil tests gpuweb/cts#2096 (1 comment by @kainino0x)
-
op: Implement 'stencil_depthFailOp_operation' test in stencil.spec.ts gpuweb/cts#2092 (1 comment by @kainino0x)
-
TR publications
-
21 documents published.
-
WebGPU published 4 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 17 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/cts
-
49 issues opened.
-
Update node dependencies gpuweb/cts#2040, by @greggman.
-
Add tests for `mix` with non-matching parameters gpuweb/cts#2039, by @zoddicus.
-
wgsl: short-named aliases for concrete vector types gpuweb/cts#2036, by @dneto0.
-
`cartesianProduct` does not correctly handle arrays of arrays for inputs gpuweb/cts#2030, by @zoddicus.
-
Add support for Matrices to Type/Value system gpuweb/cts#2029, by @zoddicus.
-
Support reading structs as outputs from builtin tests gpuweb/cts#2028, by @zoddicus.
-
Replace `new F32Interval...` with `toF32Interval` where possible gpuweb/cts#2027, by @zoddicus.
-
Use `toVector` to simplify pattern of converting from `number[]` to `Vector` gpuweb/cts#2026, by @zoddicus.
-
op: Add more tests for Depth/Stencil State gpuweb/cts#2023, by @Gyuyoung.
-
Add infinity value testing path for non-const eval numeric test gpuweb/cts#2021, by @zoddicus.
-
Remove infinity cases from numeric testing code gpuweb/cts#2020, by @zoddicus.
-
wgsl: don't reserve line, lineadj, and point gpuweb/cts#2013, by @dneto0.
-
wgsl: test abstract versions of modf and frexp gpuweb/cts#2012, by @dneto0.
-
Add CTS test for early-reject mapAsync gpuweb/cts#2009, by @takahirox.
-
wgsl: validate non-finite const eval for binary arithmetic gpuweb/cts#2008, by @amaiorano.
-
wgsl: validate non-finite const eval for smoothstep gpuweb/cts#2007, by @amaiorano.
-
wgsl: validate non-finite const eval for refract gpuweb/cts#2006, by @amaiorano.
-
wgsl: validate non-finite const eval for pack* gpuweb/cts#2005, by @amaiorano.
-
wgsl: validate non-finite const eval for normalize gpuweb/cts#2004, by @amaiorano.
-
wgsl: validate non-finite const eval for mix gpuweb/cts#2003, by @amaiorano.
-
wgsl: validate non-finite const eval for length gpuweb/cts#2002, by @amaiorano.
-
wgsl: validate non-finite const eval for fma gpuweb/cts#2001, by @amaiorano.
-
wgsl: validate non-finite const eval for dot gpuweb/cts#2000, by @amaiorano.
-
wgsl: validate non-finite const eval for cross gpuweb/cts#1999, by @amaiorano.
-
wgsl: validate non-finite const eval for clamp gpuweb/cts#1998, by @amaiorano.
-
wgsl: validate non-finite const eval for atan gpuweb/cts#1997, by @amaiorano.
-
wgsl: validate non-finite const eval for abs gpuweb/cts#1996, by @amaiorano.
-
Remove infinity cases from const eval ranges gpuweb/cts#1989, by @amaiorano.
-
Create test case to cover observed barrier miscompilation gpuweb/cts#1983, by @raphlinus.
-
rename frexp result significand field from .sig to .fract gpuweb/cts#1982, by @dneto0.
-
wgsl: & and * are lower precedence than .member and [], even in LHS gpuweb/cts#1981, by @dneto0.
-
Ability to switch input ranges when running in const vs non-const gpuweb/cts#1979, by @dj2.
-
Need to filter infinity/nan out of test cases when running with a const inputSource gpuweb/cts#1978, by @dj2.
-
op: render_pass,stencil_operation gpuweb/cts#1977, by @Gyuyoung.
-
wgsl: allow let-decl of a pointer be passed as function call argument. gpuweb/cts#1975, by @dneto0.
-
`fullF32Range` has expectation that looks suspicious gpuweb/cts#1971, by @ben-clayton.
-
op: Add more operation test for index format behavior. gpuweb/cts#1968, by @Gyuyoung.
-
wgsl: validate out of range const eval for `atanh` gpuweb/cts#1966, by @dj2.
-
wgsl: validate out of range const-eval for `asin` gpuweb/cts#1965, by @dj2.
-
wgsl: validate out of range const-eval for acos gpuweb/cts#1964, by @dj2.
-
wgsl: uniformity analysis case for short-circuiting || and && gpuweb/cts#1959, by @dneto0.
-
extractBits: don't test overflow cases for const input source gpuweb/cts#1957, by @amaiorano.
-
wgsl: brace-enclosed statement nesting depth limited to 127 gpuweb/cts#1956, by @dneto0.
-
insertBits: don't test overflow cases for const input source gpuweb/cts#1954, by @amaiorano.
-
op: rendering,draw,default_arguments gpuweb/cts#1952, by @Gyuyoung.
-
-
13 issues closed.
-
Replace `new F32Interval...` with `toF32Interval` where possible gpuweb/cts#2027.
-
wgsl: `ldexp` tests might now be incorrectly failing SwiftShader gpuweb/cts#1791.
-
wgsl: execution test for modf for vector f32 gpuweb/cts#1228.
-
wgsl: don't reserve line, lineadj, and point gpuweb/cts#2013.
-
`fullF32Range` has expectation that looks suspicious gpuweb/cts#1971.
-
op: memory_sync,buffer,ww:multiple_buffers gpuweb/cts#897.
-
wgsl: execution test for unpack2x16float builtin gpuweb/cts#1293.
-
quantizeToF16 tests do not correctly handle FTZ for f16 denorms gpuweb/cts#1946.
-
wgsl: execution test for unpack4x8snorm builtin gpuweb/cts#1289.
-
-
16 issues received 18 new comments.
-
wgsl: `ldexp` tests might now be incorrectly failing SwiftShader gpuweb/cts#1791 (1 comment by @zoddicus)
-
Replace `new F32Interval...` with `toF32Interval` where possible gpuweb/cts#2027 (1 comment by @zoddicus)
-
GPUBuffer.unmap() shouldn't fail even on unmapped or destroyed buffers gpuweb/cts#1930 (1 comment by @takahirox)
-
wgsl: function variable uniformity analysis: assignment through degenerate partial reference cuts reaching assignments gpuweb/cts#1995 (1 comment by @dneto0)
-
wgsl: & and * are lower precedence than .member and [], even in LHS gpuweb/cts#1981 (1 comment by @dneto0)
-
insertBits: don't test overflow cases for const input source gpuweb/cts#1954 (1 comment by @amaiorano)
-
42 pull requests opened.
-
update deps for various packages gpuweb/cts#2041, by @greggman.
-
Clean up a couple of .map in f32_interval.spec.ts tests gpuweb/cts#2038, by @zoddicus.
-
Replace calls to `F32Interval` constructor with `toF32Interval` gpuweb/cts#2037, by @zoddicus.
-
Fix workgroup array size limit validation test gpuweb/cts#2035, by @jrprice.
-
Only try to generate errors for testable error filters gpuweb/cts#2034, by @Kangz.
-
Test buffer.mapAsync() early rejection gpuweb/cts#2033, by @takahirox.
-
op: Implement 'stencil_compare_func' test gpuweb/cts#2032, by @Gyuyoung.
-
Handle inputs that are arrays in `cartesianProduct` gpuweb/cts#2031, by @zoddicus.
-
op: Add more parameters to 'color_write_mask,channel_work' test gpuweb/cts#2025, by @Gyuyoung.
-
op: Implement 'depth_write_disabled' test in depth.spec.ts gpuweb/cts#2024, by @Gyuyoung.
-
Fix buffer size in robust_access tests gpuweb/cts#2022, by @jrprice.
-
Expect a TypeError when transferring a mapped ArrayBuffer gpuweb/cts#2019, by @austinEng.
-
Add test case about copying HTMLVideoElement to texture gpuweb/cts#2018, by @shaoboyan.
-
Throw RangeError when creating a very large mappedAtCreation buffer gpuweb/cts#2017, by @Jiawei-Shao.
-
Remove `line`, `lineadj` and `point` from reserved words. gpuweb/cts#2015, by @dj2.
-
Implement `modf` f32 tests gpuweb/cts#2014, by @zoddicus.
-
op: Implement 'color_write_mask,channel_work' in color_target_state.spec.ts gpuweb/cts#2011, by @Gyuyoung.
-
Reduce the number of test cases generated for atan2 gpuweb/cts#2010, by @zoddicus.
-
op: Add 'color_write_mask,blending_disabled' to blending.spec.ts gpuweb/cts#1994, by @Gyuyoung.
-
gen_cache: Add `--list` mode gpuweb/cts#1993, by @ben-clayton.
-
textureSampleLevel(texture_external) -> textureSampleBaseClampToEdge() gpuweb/cts#1992, by @ben-clayton.
-
Remove infinity cases from const eval ranges gpuweb/cts#1990, by @amaiorano.
-
Implement rg11b10ufloat-renderable validation test gpuweb/cts#1988, by @takahirox.
-
operation,rendering,blending nits gpuweb/cts#1986, by @kainino0x.
-
Optimize shader expression by caching the cases gpuweb/cts#1985, by @ben-clayton.
-
op: Implement 'blend_component_clamp' test in blending.spec.ts gpuweb/cts#1976, by @Gyuyoung.
-
Simplify code around ULP calculations and fix subtly wrong tests gpuweb/cts#1974, by @zoddicus.
-
Add f16 test ranges to quantizeToF16 gpuweb/cts#1972, by @ben-clayton.
-
Fix quantizeToF16 tests to not test overflow cases for const input source gpuweb/cts#1970, by @ben-clayton.
-
Implement `fma` f32 tests gpuweb/cts#1967, by @zoddicus.
-
Convert asin, acos and atanh to a filtered range when needed. gpuweb/cts#1963, by @dj2.
-
Implement `refract` f32 tests gpuweb/cts#1962, by @zoddicus.
-
Implement 'render_bundle_commands' test in encoder_open_state.spec.ts gpuweb/cts#1961, by @Gyuyoung.
-
Fix extractBits tests to not test overflow cases for const input source gpuweb/cts#1958, by @amaiorano.
-
Fix insertBits tests to not test overflow cases for const input source gpuweb/cts#1955, by @amaiorano.
-
op: Implement 'default_arguments' test in draw.spec.ts gpuweb/cts#1953, by @Gyuyoung.
-
Add support to load cross origin resources gpuweb/cts#1951, by @shaoboyan.
-
-
44 pull requests merged.
-
Expect a TypeError when transferring a mapped ArrayBuffer gpuweb/cts#2019.
-
op: Add more parameters to 'color_write_mask,channel_work' test gpuweb/cts#2025.
-
Replace calls to `F32Interval` constructor with `toF32Interval` gpuweb/cts#2037.
-
Clean up a couple of .map in f32_interval.spec.ts tests gpuweb/cts#2038.
-
Test buffer.mapAsync() early rejection gpuweb/cts#2033.
-
Handle inputs that are arrays in `cartesianProduct` gpuweb/cts#2031.
-
Only try to generate errors for testable error filters gpuweb/cts#2034.
-
Fix workgroup array size limit validation test gpuweb/cts#2035.
-
Implement `modf` f32 tests gpuweb/cts#2014.
-
Throw RangeError when creating a very large mappedAtCreation buffer gpuweb/cts#2017.
-
Fix buffer size in robust_access tests gpuweb/cts#2022.
-
Remove `line`, `lineadj` and `point` from reserved words. gpuweb/cts#2015.
-
Reduce the number of test cases generated for atan2 gpuweb/cts#2010.
-
op: Implement 'blend_component_clamp' test in blending.spec.ts gpuweb/cts#1976.
-
op: Add 'color_write_mask,blending_disabled' to blending.spec.ts gpuweb/cts#1994.
-
Implement rg11b10ufloat-renderable validation test gpuweb/cts#1988.
-
gen_cache: Add `--list` mode gpuweb/cts#1993.
-
Remove infinity cases from const eval ranges gpuweb/cts#1990.
-
Optimize shader expression by caching the cases gpuweb/cts#1985.
-
operation,rendering,blending nits gpuweb/cts#1986.
-
Implement `refract` f32 tests gpuweb/cts#1962.
-
Implement `fma` f32 tests gpuweb/cts#1967.
-
Add rg11b10ufloat-renderable validation test plan gpuweb/cts#1937.
-
Simplify code around ULP calculations and fix subtly wrong tests gpuweb/cts#1974.
-
Add f16 test ranges to quantizeToF16 gpuweb/cts#1972.
-
Add support to load cross origin resources gpuweb/cts#1951.
-
Convert asin, acos and atanh to a filtered range when needed. gpuweb/cts#1963.
-
op: Implement 'default_arguments' test in draw.spec.ts gpuweb/cts#1953.
-
val: Implement 'pass_end_none' and 'pass_end_twice' tests gpuweb/cts#1917.
-
Implement `unpack2x16float` tests gpuweb/cts#1934.
-
Properly handle potential FTZ in quantizeToF16 tests gpuweb/cts#1947.
-
Implement `unpack4x8snorm` tests gpuweb/cts#1950.
-
-
29 pull requests received 194 new comments.
-
Add support to load cross origin resources gpuweb/cts#1951 (15 comments by @haoxli, @kainino0x, @Kangz, @shaoboyan)
-
op: Implement 'default_blend_color,initial_blend_color' test in blending.spec.ts gpuweb/cts#1980 (12 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'default_arguments' test in draw.spec.ts gpuweb/cts#1953 (12 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'blend_component_clamp' test in blending.spec.ts gpuweb/cts#1976 (11 comments by @Gyuyoung, @kainino0x)
-
Add rg11b10ufloat-renderable validation test plan gpuweb/cts#1937 (11 comments by @kainino0x, @takahirox)
-
Expect a TypeError when transferring a mapped ArrayBuffer gpuweb/cts#2019 (10 comments by @austinEng, @kainino0x)
-
op: Add 'color_write_mask,blending_disabled' to blending.spec.ts gpuweb/cts#1994 (7 comments by @Gyuyoung, @kainino0x)
-
Remove infinity cases from const eval ranges gpuweb/cts#1990 (6 comments by @amaiorano, @ben-clayton)
-
Implement rg11b10ufloat-renderable validation test gpuweb/cts#1988 (6 comments by @kainino0x, @takahirox)
-
op: Implement 'default_blend_constant,setting_blend_constant' test in blending.spec.ts gpuweb/cts#1984 (6 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'depth_write_disabled' test in depth.spec.ts gpuweb/cts#2024 (5 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'color_write_mask,channel_work' in color_target_state.spec.ts gpuweb/cts#2011 (4 comments by @Gyuyoung, @kainino0x)
-
Optimize shader expression by caching the cases gpuweb/cts#1985 (4 comments by @ben-clayton, @kainino0x)
-
Implement 'render_bundle_commands' test in encoder_open_state.spec.ts gpuweb/cts#1961 (4 comments by @Gyuyoung, @kainino0x)
-
Convert asin, acos and atanh to a filtered range when needed. gpuweb/cts#1963 (4 comments by @dj2, @zoddicus)
-
val: Introduce encoder_open_state.spec.ts to test when an encoder is open gpuweb/cts#1921 (4 comments by @kainino0x)
-
op: Implement 'simple_blend_color,blend_constant_non_inherited' test in blending.spec.ts gpuweb/cts#1987 (3 comments by @Gyuyoung, @kainino0x)
-
val: Implement 'pass_end_none' and 'pass_end_twice' tests gpuweb/cts#1917 (3 comments by @Gyuyoung, @kainino0x)
-
Fix insertBits tests to not test overflow cases for const input source gpuweb/cts#1955 (3 comments by @amaiorano, @zoddicus)
-
op: Add more parameters to 'color_write_mask,channel_work' test gpuweb/cts#2025 (2 comments by @Gyuyoung, @kainino0x)
-
Throw RangeError when creating a very large mappedAtCreation buffer gpuweb/cts#2017 (2 comments by @kainino0x)
-
Simplify code around ULP calculations and fix subtly wrong tests gpuweb/cts#1974 (2 comments by @ben-clayton, @zoddicus)
-
op: Implement 'index_format,change_pipeline_after_setIndexBuffer' test in index_format.spec.ts gpuweb/cts#1969 (2 comments by @Gyuyoung)
-
Replace calls to `F32Interval` constructor with `toF32Interval` gpuweb/cts#2037 (1 comment by @zoddicus)
gpuweb/gpuweb
-
31 issues opened.
-
Metal seem to support only up to 4096 queries in sample buffers. gpuweb/gpuweb#3648, by @Kangz.
-
Clarify sections on arithmetic divide and remainder gpuweb/gpuweb#3647, by @amaiorano.
-
Primitive clipping at the rasterization stage is inconsistent between APIs gpuweb/gpuweb#3638, by @Kangz.
-
Almost no PowerVR devices satisfy minimum Vulkan requirements gpuweb/gpuweb#3631, by @kainino0x.
-
wgsl: describe modf in terms of truncate gpuweb/gpuweb#3630, by @dneto0.
-
Suggestion to add function domains gpuweb/gpuweb#3628, by @munrocket.
-
Confusing note in builtin `length` section gpuweb/gpuweb#3621, by @amaiorano.
-
Still no way to work with simple bytes ? gpuweb/gpuweb#3620, by @dezmou.
-
[webgpu] Broken references in WebGPU gpuweb/gpuweb#3617, by @dontcallmedom-bot.
-
Missing types from reserved keywords list gpuweb/gpuweb#3613, by @teoxoy.
-
GPUBuffer memory leak gpuweb/gpuweb#3610, by @anhr.
-
WRT Vector-Matrix Multiplication and Non-Associativity gpuweb/gpuweb#3608, by @4LT.
-
wgsl: full reference to module-scope variable is uniform, the value stored there may not be gpuweb/gpuweb#3602, by @dneto0.
-
Optional feature request: filterable 32-bit float textures gpuweb/gpuweb#3601, by @magcius.
-
destroy() should unmap if there's a pending mapping gpuweb/gpuweb#3597, by @kainino0x.
-
GPUBuffer.mapAsync is very slow... gpuweb/gpuweb#3595, by @tlecoz.
-
Naming convention for sync and async APIs gpuweb/gpuweb#3594, by @gyagp.
-
clarify const-expression short-circuiting operators gpuweb/gpuweb#3591, by @dneto0.
-
bitwise operator not working anymore ? gpuweb/gpuweb#3590, by @bzhmaddog.
-
Basic questions about ComputeShader gpuweb/gpuweb#3584, by @tlecoz.
-
Impact of WGSL Uniformity Check Compatibility on Performance gpuweb/gpuweb#3583, by @hana-alice.
-
wgsl: "constant identifier expression" section needs to adapt for `const` gpuweb/gpuweb#3578, by @dneto0.
-
WebGPU: Forbid null code point '\0' (U+0000) within strings gpuweb/gpuweb#3576, by @jiangzhaoming.
-
wgsl: fp conversion overflow is error when const-expr (not just abstract type) gpuweb/gpuweb#3571, by @dneto0.
-
wgsl: consider renaming frexp result fields gpuweb/gpuweb#3569, by @dneto0.
-
wgsl: explain the mnemonic for fma, ldexp, frexp gpuweb/gpuweb#3567, by @dneto0.
-
rg11b10ufloat-renderable should probably allow use as a resolve target gpuweb/gpuweb#3566, by @kainino0x.
-
33 issues closed.
-
Add precedence rules description into spec gpuweb/gpuweb#3346.
-
Still no way to work with simple bytes ? gpuweb/gpuweb#3620.
-
GPUBuffer memory leak gpuweb/gpuweb#3610.
-
vec4<f32> is too hard to type on the keyboard gpuweb/gpuweb#736.
-
wgsl: describe modf in terms of truncate gpuweb/gpuweb#3630.
-
Suggestion to add function domains gpuweb/gpuweb#3628.
-
WRT Vector-Matrix Multiplication and Non-Associativity gpuweb/gpuweb#3608.
-
add AbstractFloat overloads for modf and frexp? gpuweb/gpuweb#3358.
-
why doesn't WGSL have mediump floats gpuweb/gpuweb#3524.
-
Naming convention for sync and async APIs gpuweb/gpuweb#3594.
-
Missing types from reserved keywords list gpuweb/gpuweb#3613.
-
access decorations for textures (text/example inconsistency) gpuweb/gpuweb#1419.
-
wgsl editorial: uniformity rules say "reference to ...." and that's confusing gpuweb/gpuweb#3457.
-
Consider limiting the set of override declarations in shader stage interface gpuweb/gpuweb#3555.
-
destroy() should unmap if there's a pending mapping gpuweb/gpuweb#3597.
-
GPUBuffer.mapAsync is very slow... gpuweb/gpuweb#3595.
-
bitwise operator not working anymore ? gpuweb/gpuweb#3590.
-
clarify const-expression short-circuiting operators gpuweb/gpuweb#3591.
-
Impact of WGSL Uniformity Check Compatibility on Performance gpuweb/gpuweb#3583.
-
wgsl: consider renaming frexp result fields gpuweb/gpuweb#3569.
-
Runtime behavior of NaN-producing math functions gpuweb/gpuweb#3365.
-
wgsl: Add limits to avoid nuisance floating point literals gpuweb/gpuweb#3255.
-
Basic questions about ComputeShader gpuweb/gpuweb#3584.
-
wgsl: "constant identifier expression" section needs to adapt for `const` gpuweb/gpuweb#3578.
-
wgsl: explain the mnemonic for fma, ldexp, frexp gpuweb/gpuweb#3567.
-
wgsl: fp conversion overflow is error when const-expr (not just abstract type) gpuweb/gpuweb#3571.
-
wgsl: limit nesting depth of if-else gpuweb/gpuweb#3527.
-
Shader stage input-output subsetting rules gpuweb/gpuweb#2038.
-
WGSL pointer aliasing rules gpuweb/gpuweb#1457.
-
-
53 issues received 173 new comments.
-
Designing a uniformity opt-out gpuweb/gpuweb#3554 (21 comments by @alan-baker, @dneto0, @fintelia, @jrprice, @kdashg, @krogovin, @RSSchermer, @tex3d)
-
wgsl: add way to get special values for a given numeric type (e.g. max, min, lowest) gpuweb/gpuweb#3431 (14 comments by @ben-clayton, @dneto0, @kdashg)
-
Workgroup Broadcast gpuweb/gpuweb#2586 (12 comments by @alan-baker, @dneto0, @kdashg, @litherum, @raphlinus)
-
WebGPU: Forbid null code point '\0' (U+0000) within strings gpuweb/gpuweb#3576 (9 comments by @greggman, @jiangzhaoming, @kainino0x, @Kangz, @kdashg, @toji)
-
GPUBuffer memory leak gpuweb/gpuweb#3610 (6 comments by @anhr, @ben-clayton, @kainino0x, @kdashg, @munrocket)
-
How do we specify the timing of external texture expiration? gpuweb/gpuweb#3324 (6 comments by @kainino0x, @kdashg, @shaoboyan, @toji)
-
Primitive clipping at the rasterization stage is inconsistent between APIs gpuweb/gpuweb#3638 (5 comments by @austinEng, @kainino0x, @Kangz)
-
Optional feature request: filterable 32-bit float textures gpuweb/gpuweb#3601 (4 comments by @kainino0x, @magcius)
-
Naming convention for sync and async APIs gpuweb/gpuweb#3594 (4 comments by @gyagp, @kainino0x, @kdashg)
-
Almost no PowerVR devices satisfy minimum Vulkan requirements gpuweb/gpuweb#3631 (4 comments by @kainino0x)
-
WRT Vector-Matrix Multiplication and Non-Associativity gpuweb/gpuweb#3608 (4 comments by @4LT, @kainino0x, @mehmetoguzderin)
-
Missing types from reserved keywords list gpuweb/gpuweb#3613 (4 comments by @ben-clayton, @munrocket, @teoxoy)
-
wgsl: full reference to module-scope variable is uniform, the value stored there may not be gpuweb/gpuweb#3602 (4 comments by @dneto0)
-
GPUBuffer.mapAsync is very slow... gpuweb/gpuweb#3595 (4 comments by @ben-clayton, @kainino0x, @tlecoz)
-
Multiple components cannot effectively share a GPUDevice because it is stateful gpuweb/gpuweb#3250 (4 comments by @kainino0x, @kdashg)
-
bitwise operator not working anymore ? gpuweb/gpuweb#3590 (4 comments by @alan-baker, @ben-clayton, @bzhmaddog)
-
Impact of WGSL Uniformity Check Compatibility on Performance gpuweb/gpuweb#3583 (4 comments by @dneto0, @hana-alice, @jiangzhaoming)
-
Runtime behavior of NaN-producing math functions gpuweb/gpuweb#3365 (4 comments by @dneto0, @kdashg)
-
Should "mark adapters stale" be called more quickly? gpuweb/gpuweb#1630 (3 comments by @kainino0x, @kdashg)
-
rg11b10ufloat-renderable should probably allow use as a resolve target gpuweb/gpuweb#3566 (3 comments by @kainino0x, @kdashg, @toji)
-
consider expanding concept of full reference: include indexing into single-element array, and member access of single-member struct? gpuweb/gpuweb#3605 (3 comments by @dneto0, @kdashg)
-
Define domain for float builtins and describe errors gpuweb/gpuweb#3220 (2 comments by @dj2, @dneto0)
-
Consider limiting the set of override declarations in shader stage interface gpuweb/gpuweb#3555 (2 comments by @dneto0, @kdashg)
-
Accessing an "invalid memory reference": Trapping behavior gpuweb/gpuweb#3272 (2 comments by @kdashg)
-
WebGL - WebGPU interoperability #18 "copy interaction" - what about an easy post process option gpuweb/gpuweb#3574 (2 comments by @dobkeratops, @josephrocca)
-
Improve color and display-size readability of diagrams and code blocks and other elements in WGSL spec gpuweb/gpuweb#3649 (1 comment by @kainino0x)
-
Buffer Offset issue in buffer-texture copies with depth stencil textures gpuweb/gpuweb#3185 (1 comment by @lokokung)
-
Behavior of endPass() and pass commands after pass has ended gpuweb/gpuweb#2464 (1 comment by @lokokung)
-
Definition of length(x) is too restrictive for x being a scalar gpuweb/gpuweb#3407 (1 comment by @munrocket)
-
access decorations for textures (text/example inconsistency) gpuweb/gpuweb#1419 (1 comment by @jrprice)
-
wgsl editorial: uniformity rules say "reference to ...." and that's confusing gpuweb/gpuweb#3457 (1 comment by @dneto0)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (1 comment by @aphillips)
-
Should the API enforce that writable storage buffer bindings don't alias each other? gpuweb/gpuweb#1842 (1 comment by @dneto0)
-
Allow signedness to differ between integer texture builtin arguments (`C` template) gpuweb/gpuweb#3536 (1 comment by @kdashg)
-
wgsl: grammar for LHS is ambiguous about grouping its subexpressions, affects type checking gpuweb/gpuweb#3530 (1 comment by @kdashg)
-
-
51 pull requests opened.
-
Reduce query set count limit to 4096 gpuweb/gpuweb#3650, by @kainino0x.
-
wgsl: clarify function domains gpuweb/gpuweb#3646, by @munrocket.
-
[editorial] Check all destroyable resources at submit() gpuweb/gpuweb#3645, by @kainino0x.
-
Add uniformity analysis opt-out for derivative-using builtins gpuweb/gpuweb#3644, by @dneto0.
-
Fix typo in compute pipeline creation section gpuweb/gpuweb#3643, by @jrprice.
-
WGSL: allow bgra8unorm as texel format gpuweb/gpuweb#3640, by @Jiawei-Shao.
-
[editorial] Introduce behavioral requirement definition gpuweb/gpuweb#3639, by @alan-baker.
-
wgsl: sample `texture_external` with `textureSampleBaseClampToEdge` gpuweb/gpuweb#3637, by @jiangzhaoming.
-
Fix typos gpuweb/gpuweb#3636, by @mehmetoguzderin.
-
Describe modf in terms of trunc and subtraction gpuweb/gpuweb#3635, by @dneto0.
-
Use automatic expiry for imported textures, remove .expired gpuweb/gpuweb#3633, by @kainino0x.
-
Use task sources for device events and canvas texture expiry gpuweb/gpuweb#3632, by @kainino0x.
-
wgsl: Make builtin returned struct names consistent gpuweb/gpuweb#3629, by @ben-clayton.
-
Clarify "to WGSL type" use of TypeError gpuweb/gpuweb#3627, by @kainino0x.
-
Fix issues in pass end() validation gpuweb/gpuweb#3626, by @kainino0x.
-
wgsl: Fix typo in `modf` description gpuweb/gpuweb#3625, by @dneto0.
-
wgsl: Fix typo .sig -> .fract gpuweb/gpuweb#3624, by @munrocket.
-
Add abstract versions of modf and frexp gpuweb/gpuweb#3623, by @alan-baker.
-
[editorial] Remove note about non-meaningful result of scalar length gpuweb/gpuweb#3622, by @alan-baker.
-
wgsl: Fix typo in `modf` description gpuweb/gpuweb#3619, by @ben-clayton.
-
wgsl: Don't reserve line, lineadj, point gpuweb/gpuweb#3618, by @dneto0.
-
wgsl: Don't reserve line, lineadj, point gpuweb/gpuweb#3615, by @munrocket.
-
Fix capitalization of section title gpuweb/gpuweb#3614, by @mehmetoguzderin.
-
Introduce Operator Precedence and Associativity gpuweb/gpuweb#3609, by @mehmetoguzderin.
-
wgsl: Capitalize the first word after "Note:" gpuweb/gpuweb#3607, by @dneto0.
-
wgsl: describe partial references that still cover all memory locations gpuweb/gpuweb#3606, by @dneto0.
-
wgsl: Make sections for each barrier function gpuweb/gpuweb#3604, by @dneto0.
-
Fix confusing "reference to" wording in uniformity analysis gpuweb/gpuweb#3603, by @dneto0.
-
wgsl: Limit the overrides declarations in shader interface gpuweb/gpuweb#3600, by @dneto0.
-
Refactor argument_expression_list for readability gpuweb/gpuweb#3599, by @dneto0.
-
GPUBuffer.destroy() calls unmap() if pending gpuweb/gpuweb#3598, by @takahirox.
-
Remove early statement saying variables don't overlap gpuweb/gpuweb#3596, by @dneto0.
-
Fix typo (const-expression -> override-expression) gpuweb/gpuweb#3593, by @terier.
-
clarify when const- and override-expressions are evaluated gpuweb/gpuweb#3592, by @dneto0.
-
Fix description of fract gpuweb/gpuweb#3589, by @dneto0.
-
Rename frexp result significand to .fract gpuweb/gpuweb#3588, by @dneto0.
-
Add workgroupuniform load built-in and analysis gpuweb/gpuweb#3586, by @alan-baker.
-
Remove 0 return when a NaN would be produced for atanh and acosh gpuweb/gpuweb#3585, by @alan-baker.
-
Disambiguate precedence of & and * vs. [] and .member in lhs_expression gpuweb/gpuweb#3582, by @dneto0.
-
Allow aliased resource variables gpuweb/gpuweb#3581, by @alan-baker.
-
Explicitly define a root identifier gpuweb/gpuweb#3580, by @dneto0.
-
Update identifier expressions table to include const-decl gpuweb/gpuweb#3579, by @dneto0.
-
Allow slightly more expressions for pointer arguments gpuweb/gpuweb#3577, by @alan-baker.
-
Add samples gpuweb/gpuweb#3573, by @pixcai.
-
float conversion overflow is compile-time error if orginal expression… gpuweb/gpuweb#3572, by @dneto0.
-
Add mnemonics for ldexp and fma gpuweb/gpuweb#3570, by @dneto0.
-
Describe a mnemonic for 'frexp' gpuweb/gpuweb#3568, by @dneto0.
-
Clarify mathematical value of float literal gpuweb/gpuweb#3565, by @dneto0.
-
wgsl: add compile-time failure cases for quantizeToF16, pack2x16float gpuweb/gpuweb#3564, by @dneto0.
-
Clarify rule about binding aliasing in a shader gpuweb/gpuweb#3562, by @dneto0.
-
Document extreme values for concrete numeric scalar types gpuweb/gpuweb#3561, by @dneto0.
-
-
44 pull requests merged.
-
[editorial] Check all destroyable resources at submit() gpuweb/gpuweb#3645.
-
Clarify "to WGSL type" use of TypeError gpuweb/gpuweb#3627.
-
Introduce Operator Precedence and Associativity gpuweb/gpuweb#3609.
-
wgsl: sample `texture_external` with `textureSampleBaseClampToEdge` gpuweb/gpuweb#3637.
-
Fix typo in compute pipeline creation section gpuweb/gpuweb#3643.
-
[editorial] Introduce behavioral requirement definition gpuweb/gpuweb#3639.
-
Add predeclared short names for vector types gpuweb/gpuweb#3291.
-
Describe modf in terms of trunc and subtraction gpuweb/gpuweb#3635.
-
Fix typos gpuweb/gpuweb#3636.
-
wgsl: describe partial references that still cover all memory locations gpuweb/gpuweb#3606.
-
wgsl: Make builtin returned struct names consistent gpuweb/gpuweb#3629.
-
wgsl: Fix typo in `modf` description gpuweb/gpuweb#3625.
-
wgsl: Fix typo .sig -> .fract gpuweb/gpuweb#3624.
-
Add abstract versions of modf and frexp gpuweb/gpuweb#3623.
-
wgsl: Don't reserve line, lineadj, point gpuweb/gpuweb#3615.
-
Fix capitalization of section title gpuweb/gpuweb#3614.
-
wgsl: Capitalize the first word after "Note:" gpuweb/gpuweb#3607.
-
wgsl: Make sections for each barrier function gpuweb/gpuweb#3604.
-
Fix confusing "reference to" wording in uniformity analysis gpuweb/gpuweb#3603.
-
wgsl: Limit the overrides declarations in shader interface gpuweb/gpuweb#3600.
-
GPUBuffer.destroy() calls unmap() if pending gpuweb/gpuweb#3598.
-
Refactor argument_expression_list for readability gpuweb/gpuweb#3599.
-
[editorial] Make a section for defining copies generally, not just commands gpuweb/gpuweb#3519.
-
Remove early statement saying variables don't overlap gpuweb/gpuweb#3596.
-
Fix typo (const-expression -> override-expression) gpuweb/gpuweb#3593.
-
clarify when const- and override-expressions are evaluated gpuweb/gpuweb#3592.
-
Fix description of fract gpuweb/gpuweb#3589.
-
Rename frexp result significand to .fract gpuweb/gpuweb#3588.
-
Disambiguate precedence of & and * vs. [] and .member in lhs_expression gpuweb/gpuweb#3582.
-
Remove 0 return when a NaN would be produced for atanh and acosh gpuweb/gpuweb#3585.
-
Clarify mathematical value of float literal gpuweb/gpuweb#3565.
-
Explicitly define a root identifier gpuweb/gpuweb#3580.
-
Allow aliased resource variables gpuweb/gpuweb#3581.
-
Update identifier expressions table to include const-decl gpuweb/gpuweb#3579.
-
Add mnemonics for ldexp and fma gpuweb/gpuweb#3570.
-
Describe a mnemonic for 'frexp' gpuweb/gpuweb#3568.
-
Allow slightly more expressions for pointer arguments gpuweb/gpuweb#3577.
-
float conversion overflow is compile-time error if orginal expression… gpuweb/gpuweb#3572.
-
Clarify rule about binding aliasing in a shader gpuweb/gpuweb#3562.
-
wgsl: add compile-time failure cases for quantizeToF16, pack2x16float gpuweb/gpuweb#3564.
-
Remove special-case uniformity rules for short-circuiting ops gpuweb/gpuweb#3558.
-
Document extreme values for concrete numeric scalar types gpuweb/gpuweb#3561.
-
Limit the maximum statement nesting depth to 127 gpuweb/gpuweb#3552.
-
Add compile/pipeline-time error cases for insertBits gpuweb/gpuweb#3560.
-
-
6 pull requests closed.
-
Remove non-generic types from keywords list. gpuweb/gpuweb#3285.
-
Introduce Subgroup Operations Extension gpuweb/gpuweb#1459.
-
wgsl: Fix typo in `modf` description gpuweb/gpuweb#3619.
-
wgsl: Don't reserve line, lineadj, point gpuweb/gpuweb#3618.
-
mapAsync(WRITE) and zero-filling gpuweb/gpuweb#2926.
-
wgsl: Move * and & from lhs_expression to core_lhs_expression gpuweb/gpuweb#3531.
-
-
52 pull requests received 207 new comments.
-
Introduce Operator Precedence and Associativity gpuweb/gpuweb#3609 (34 comments by @alan-baker, @kainino0x, @mehmetoguzderin)
-
Add workgroupuniform load built-in and analysis gpuweb/gpuweb#3586 (19 comments by @alan-baker, @dneto0, @jrprice, @Kangz)
-
WGSL syntax styling upgrade with initial notation section gpuweb/gpuweb#3543 (17 comments by @dj2, @dneto0, @mehmetoguzderin)
-
Add uniformity analysis opt-out for derivative-using builtins gpuweb/gpuweb#3644 (14 comments by @alan-baker, @dneto0, @jimblandy, @jrprice, @Kangz, @mehmetoguzderin)
-
Clarify mathematical value of float literal gpuweb/gpuweb#3565 (11 comments by @dj2, @dneto0, @kdashg)
-
Allow slightly more expressions for pointer arguments gpuweb/gpuweb#3577 (11 comments by @alan-baker, @ben-clayton, @dneto0)
-
wgsl: describe partial references that still cover all memory locations gpuweb/gpuweb#3606 (7 comments by @alan-baker, @dneto0)
-
[editorial] Introduce behavioral requirement definition gpuweb/gpuweb#3639 (6 comments by @alan-baker, @dneto0)
-
wgsl: Make builtin returned struct names consistent gpuweb/gpuweb#3629 (5 comments by @dneto0, @mehmetoguzderin)
-
wgsl: Limit the overrides declarations in shader interface gpuweb/gpuweb#3600 (5 comments by @dneto0, @jrprice)
-
wgsl: Fix typo in `modf` description gpuweb/gpuweb#3619 (4 comments by @ben-clayton, @dneto0, @munrocket)
-
GPUBuffer.destroy() calls unmap() if pending gpuweb/gpuweb#3598 (4 comments by @kainino0x, @takahirox)
-
Use automatic expiry for imported textures, remove .expired gpuweb/gpuweb#3633 (3 comments by @toji)
-
Remove early statement saying variables don't overlap gpuweb/gpuweb#3596 (3 comments by @dneto0, @mehmetoguzderin)
-
clarify when const- and override-expressions are evaluated gpuweb/gpuweb#3592 (3 comments by @alan-baker, @dneto0)
-
Rename frexp result significand to .fract gpuweb/gpuweb#3588 (3 comments by @dneto0, @mehmetoguzderin)
-
Document extreme values for concrete numeric scalar types gpuweb/gpuweb#3561 (3 comments by @dneto0)
-
Reduce max binding index and make it a limit gpuweb/gpuweb#3318 (2 comments by @kainino0x, @lokokung)
-
Use task sources for device events and canvas texture expiry gpuweb/gpuweb#3632 (2 comments by @toji)
-
Add HTMLImageElement and ImageData to copyExternalImageToTexture gpuweb/gpuweb#3430 (2 comments by @kdashg)
-
Require RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/gpuweb#2786 (2 comments by @kainino0x, @takahirox)
-
Disambiguate precedence of & and * vs. [] and .member in lhs_expression gpuweb/gpuweb#3582 (2 comments by @dneto0, @mehmetoguzderin)
-
Allow using depth textures with 'unfilterable-float' bindings gpuweb/gpuweb#3230 (1 comment by @lokokung)
-
Allow BGL entries for multisampled:true "depth" textures gpuweb/gpuweb#3072 (1 comment by @lokokung)
-
Fix confusing "reference to" wording in uniformity analysis gpuweb/gpuweb#3603 (1 comment by @dneto0)
-
wgsl: Move * and & from lhs_expression to core_lhs_expression gpuweb/gpuweb#3531 (1 comment by @dneto0)
-
float conversion overflow is compile-time error if orginal expression… gpuweb/gpuweb#3572 (1 comment by @dneto0)
-
Add draft text/wgsl media type registration. WGSL uses UTF-8 gpuweb/gpuweb#3542 (1 comment by @kdashg)
-
TR publications
-
17 documents published.
-
WebGPU Shading Language published 9 times as a Working Draft. History of WebGPU Shading Language
-
WebGPU published 8 times as a Working Draft. History of WebGPU
-
GitHub
gpuweb/gpuweb
-
29 issues opened.
-
support blending 32 bit float textures gpuweb/gpuweb#3556, by @mighdoll.
-
Consider limiting the set of override declarations in shader stage interface gpuweb/gpuweb#3555, by @jrprice.
-
Designing a uniformity opt-out gpuweb/gpuweb#3554, by @alan-baker.
-
Does the spec need to mention \0 in strings? gpuweb/gpuweb#3550, by @greggman.
-
Support fuller-featured pointers (as per #3276) gpuweb/gpuweb#3546, by @kdashg.
-
About destroy and memory relationship gpuweb/gpuweb#3537, by @redcamel.
-
Allow signedness to differ between integer texture builtin arguments (`C` template) gpuweb/gpuweb#3536, by @ben-clayton.
-
Fixed-footprint note can be shortened gpuweb/gpuweb#3534, by @ben-clayton.
-
wgsl: NaNs and infinities in uniform or storage buffer gpuweb/gpuweb#3533, by @jiangzhaoming.
-
wgsl: grammar for LHS is ambiguous about grouping its subexpressions, affects type checking gpuweb/gpuweb#3530, by @dneto0.
-
wgsl: rename grammar rules "postfix_expression" because it's not an expression gpuweb/gpuweb#3528, by @dneto0.
-
wgsl: limit nesting depth of if-else gpuweb/gpuweb#3527, by @dneto0.
-
why doesn't WGSL have mediump floats gpuweb/gpuweb#3524, by @dneto0.
-
Attribute syntax should not specialize to built-in language attributes gpuweb/gpuweb#3520, by @mehmetoguzderin.
-
wgsl: address space names should be context-dependent names instead of keywords gpuweb/gpuweb#3517, by @dneto0.
-
[editorial] Improve diagrams for buffer mapping (and other timeline things) gpuweb/gpuweb#3513, by @kainino0x.
-
wgsl: (1 << 2u) should be 4 (an abstract int) gpuweb/gpuweb#3511, by @dneto0.
-
Consider removing the loop statement gpuweb/gpuweb#3504, by @munrocket.
-
Make WGSL more suitable for minification gpuweb/gpuweb#3503, by @munrocket.
-
Spec doesn't allow local variables with the same names as parameters gpuweb/gpuweb#3502, by @jimblandy.
-
Tuple-like-returns proposal for function-out-params functionality gpuweb/gpuweb#3501, by @kdashg.
-
wgsl: Flat interpolants are uniform across fragments in a fragment shader invocation group gpuweb/gpuweb#3489, by @dneto0.
-
Unclear whether GPUExternalTexture is 1D, 2D or 3D gpuweb/gpuweb#3487, by @i509VCB.
-
Global Sync gpuweb/gpuweb#3486, by @m-schuetz.
-
Validating maxComputeWorkgroupStorageSize gpuweb/gpuweb#3485, by @kainino0x.
-
Permissions Policy for WebGPU and for unmaskHints gpuweb/gpuweb#3483, by @kainino0x.
-
32 issues closed.
-
[wgsl] Encoding for shader text needs to be specified gpuweb/gpuweb#565.
-
wgsl: NaNs and infinities in uniform or storage buffer gpuweb/gpuweb#3533.
-
Does the spec need to mention \0 in strings? gpuweb/gpuweb#3550.
-
Unclear whether GPUExternalTexture is 1D, 2D or 3D gpuweb/gpuweb#3487.
-
wgsl: pointer-to-workgroup in user-defined functions gpuweb/gpuweb#3276.
-
Defining "Undefined Value" gpuweb/gpuweb#2888.
-
Fixed-footprint note can be shortened gpuweb/gpuweb#3534.
-
About destroy and memory relationship gpuweb/gpuweb#3537.
-
wgsl: rename grammar rules "postfix_expression" because it's not an expression gpuweb/gpuweb#3528.
-
Validating maxComputeWorkgroupStorageSize gpuweb/gpuweb#3485.
-
wgsl: (1 << 2u) should be 4 (an abstract int) gpuweb/gpuweb#3511.
-
Difference between `src-alpha` and `src-alpha-saturated` gpuweb/gpuweb#3432.
-
Make WGSL more suitable for minification gpuweb/gpuweb#3503.
-
Pointers in uniformity analysis gpuweb/gpuweb#2741.
-
Error with literal indexing in a shader gpuweb/gpuweb#1803.
-
wgsl: can hex floats explicitly represent infinity and NaN as float literals? gpuweb/gpuweb#1769.
-
wgsl: address space names should be context-dependent names instead of keywords gpuweb/gpuweb#3517.
-
Clarify intended behaviour when packing out of range values gpuweb/gpuweb#3470.
-
Consider removing the loop statement gpuweb/gpuweb#3504.
-
wgsl: trunc behaviour is not specified for infinity gpuweb/gpuweb#3146.
-
Fragment output <-> attachment interface matching rules def gpuweb/gpuweb#2013.
-
Object labels require client-shared state gpuweb/gpuweb#3263.
-
Does getBindGroupLayout return the same underlying object or a new one? gpuweb/gpuweb#3454.
-
Bitcasting to NaN should be a compile error? gpuweb/gpuweb#3468.
-
Global Sync gpuweb/gpuweb#3486.
-
wgsl: atan2 result depends on signs of both arguments, not spec'd gpuweb/gpuweb#3437.
-
-
3 issues reopened.
-
Define a MIME type gpuweb/gpuweb#1682, by @dneto0.
-
Should the API enforce that writable storage buffer bindings don't alias each other? gpuweb/gpuweb#1842, by @kainino0x.
-
Shader stage input-output subsetting rules gpuweb/gpuweb#2038, by @kainino0x.
-
-
45 issues received 174 new comments.
-
Uniformity analysis is far too strict gpuweb/gpuweb#3479 (33 comments by @alan-baker, @ben-clayton, @brendan-duncan, @dneto0, @fintelia, @jimblandy, @jrprice, @Kangz, @kdashg, @krogovin, @raphlinus)
-
Make WGSL more suitable for minification gpuweb/gpuweb#3503 (12 comments by @dj2, @dneto0, @mehmetoguzderin, @munrocket)
-
Spec doesn't allow local variables with the same names as parameters gpuweb/gpuweb#3502 (10 comments by @dj2, @dneto0, @jimblandy, @kdashg, @mehmetoguzderin)
-
Canvas texture expiry timing is unpredictable gpuweb/gpuweb#3295 (8 comments by @greggman, @kainino0x, @kdashg, @litherum, @toji)
-
Tuple-like-returns proposal for function-out-params functionality gpuweb/gpuweb#3501 (8 comments by @ben-clayton, @dneto0, @litherum, @mehmetoguzderin, @munrocket)
-
createPipelineAsync() doesn't reject its promise, even if the created pipeline is invalid gpuweb/gpuweb#3296 (7 comments by @kainino0x, @kdashg, @litherum)
-
create*PipelineAsync prevents using the pipeline synchronously gpuweb/gpuweb#3297 (7 comments by @greggman, @kainino0x, @magcius)
-
wgsl: add way to get special values for a given numeric type (e.g. max, min, lowest) gpuweb/gpuweb#3431 (7 comments by @dj2, @dneto0, @kdashg, @litherum)
-
wgsl: grammar for LHS is ambiguous about grouping its subexpressions, affects type checking gpuweb/gpuweb#3530 (6 comments by @djg, @dneto0, @jimblandy, @kdashg, @mehmetoguzderin)
-
Does the spec need to mention \0 in strings? gpuweb/gpuweb#3550 (5 comments by @ben-clayton, @greggman, @Kangz, @kdashg)
-
Fixed-footprint note can be shortened gpuweb/gpuweb#3534 (5 comments by @ben-clayton, @dj2, @dneto0)
-
Allow signedness to differ between integer texture builtin arguments (`C` template) gpuweb/gpuweb#3536 (4 comments by @ben-clayton, @dneto0)
-
wgsl: pointer-to-workgroup in user-defined functions gpuweb/gpuweb#3276 (4 comments by @kdashg, @litherum)
-
Shift left for concrete: "e2+1 most significant bits of e1 do not have the same bit value" should only apply to signed values (i32) gpuweb/gpuweb#3535 (4 comments by @amaiorano, @dneto0)
-
Validating maxComputeWorkgroupStorageSize gpuweb/gpuweb#3485 (4 comments by @dneto0, @Kangz, @kdashg)
-
support blending 32 bit float textures gpuweb/gpuweb#3556 (3 comments by @kainino0x, @kdashg, @mighdoll)
-
wgsl: limit nesting depth of if-else gpuweb/gpuweb#3527 (3 comments by @ben-clayton, @dneto0, @kdashg)
-
Clarify behavior of float division/remainder by 0 gpuweb/gpuweb#2798 (3 comments by @djg, @dneto0, @kdashg)
-
How do we specify the timing of external texture expiration? gpuweb/gpuweb#3324 (2 comments by @kainino0x, @Kangz)
-
MSL does not support address-of vector component gpuweb/gpuweb#1931 (2 comments by @dneto0, @jrprice)
-
invalid expressions in const-expr and override-expr should be error (no later than pipeline creation time) gpuweb/gpuweb#3253 (2 comments by @dneto0, @kdashg)
-
wgsl: (1 << 2u) should be 4 (an abstract int) gpuweb/gpuweb#3511 (2 comments by @ben-clayton, @dneto0)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (2 comments by @kainino0x, @kdashg)
-
Unclear whether GPUExternalTexture is 1D, 2D or 3D gpuweb/gpuweb#3487 (2 comments by @kainino0x, @Kangz)
-
Should the API enforce that writable storage buffer bindings don't alias each other? gpuweb/gpuweb#1842 (1 comment by @kainino0x)
-
wgsl: NaNs and infinities in uniform or storage buffer gpuweb/gpuweb#3533 (1 comment by @alan-baker)
-
wgsl: spec still says you can't declare something at module-scope to shadow a built-in function gpuweb/gpuweb#3525 (1 comment by @dneto0)
-
Attribute syntax should not specialize to built-in language attributes gpuweb/gpuweb#3520 (1 comment by @mehmetoguzderin)
-
wgsl: can hex floats explicitly represent infinity and NaN as float literals? gpuweb/gpuweb#1769 (1 comment by @dneto0)
-
wgsl: address space names should be context-dependent names instead of keywords gpuweb/gpuweb#3517 (1 comment by @dneto0)
-
wgsl: Flat interpolants are uniform across fragments in a fragment shader invocation group gpuweb/gpuweb#3489 (1 comment by @alan-baker)
-
-
46 pull requests opened.
-
Add compile/pipeline-time error cases for insertBits gpuweb/gpuweb#3560, by @dneto0.
-
Remove special-case uniformity rules for short-circuiting ops gpuweb/gpuweb#3558, by @jrprice.
-
Say "indeterminate value" where previously said "undefined value" gpuweb/gpuweb#3557, by @dneto0.
-
Fix clamp definition gpuweb/gpuweb#3553, by @alan-baker.
-
Limit the maximum statement nesting depth to 127 gpuweb/gpuweb#3552, by @alan-baker.
-
Update samples to the latest spec gpuweb/gpuweb#3551, by @pixcai.
-
[editorial] Clarify nan/inf eval applies to all expressions gpuweb/gpuweb#3549, by @alan-baker.
-
Generalize integer texture parameter requirements gpuweb/gpuweb#3548, by @alan-baker.
-
Force concrete types unless all subexpressions are const-expressions gpuweb/gpuweb#3547, by @alan-baker.
-
Specific GPUExternalTexture dimension in spec gpuweb/gpuweb#3544, by @shaoboyan.
-
WGSL syntax styling upgrade with initial notation section gpuweb/gpuweb#3543, by @mehmetoguzderin.
-
Add draft text/wgsl media type registration. WGSL uses UTF-8 gpuweb/gpuweb#3542, by @dneto0.
-
wgsl: Simplify note about runtime-sized array not being fixed-footprint gpuweb/gpuweb#3540, by @dneto0.
-
wgsl: Fix how we say concrete left shift must not overflow (if known … gpuweb/gpuweb#3539, by @dneto0.
-
Set GPUBuffer.[[pending_map]] to null when mapAsync() succeeds gpuweb/gpuweb#3532, by @takahirox.
-
wgsl: Move * and & from lhs_expression to core_lhs_expression gpuweb/gpuweb#3531, by @dneto0.
-
Rename postfix_expression to component_or_swizzle_specifier gpuweb/gpuweb#3529, by @dneto0.
-
Remove stale/incorrect statement about inability to shadow builtins gpuweb/gpuweb#3526, by @dneto0.
-
Reject create*PipelineAsync on failure gpuweb/gpuweb#3523, by @kainino0x.
-
workgroupStorageUsed is sum of values, each rounded up to multiple of 16 gpuweb/gpuweb#3522, by @dneto0.
-
Fix the definition of "src-alpha-saturated" gpuweb/gpuweb#3521, by @Jiawei-Shao.
-
[editorial] Make a section for defining copies generally, not just commands gpuweb/gpuweb#3519, by @kainino0x.
-
wgsl: Remove address space names from "other keywords" list gpuweb/gpuweb#3518, by @dneto0.
-
wgsl: bit shift width is always u32 or vector of u32 gpuweb/gpuweb#3516, by @dneto0.
-
[editorial] Misc trivial inline issue cleanup gpuweb/gpuweb#3515, by @kainino0x.
-
[editorial] Consolidate "Detailed Operations" issues into one issue gpuweb/gpuweb#3514, by @kainino0x.
-
[editorial] Move texture extent algorithms to texture section gpuweb/gpuweb#3512, by @kainino0x.
-
[editorial] Describe textures and subresource accounting algorithms gpuweb/gpuweb#3510, by @kainino0x.
-
[editorial] Clarify some spec around features gpuweb/gpuweb#3509, by @kainino0x.
-
wgsl: Define "indeterminate value" gpuweb/gpuweb#3508, by @dneto0.
-
Clarify behaviour of pack2x16float, quantizeToF16 for large inputs gpuweb/gpuweb#3507, by @dneto0.
-
Don't throw exception when getBindGroupLayout index > maxBindGroups gpuweb/gpuweb#3506, by @kainino0x.
-
Don't trail with empty BGLs in default pipeline layout gpuweb/gpuweb#3505, by @kainino0x.
-
Only allow defined GPUShaderStage bits in bind group layout entries gpuweb/gpuweb#3500, by @kainino0x.
-
Only allow defined GPUMapMode bits in mapAsync() gpuweb/gpuweb#3499, by @kainino0x.
-
[editorial] Replace "super-luminant" with a more descriptive term gpuweb/gpuweb#3498, by @kainino0x.
-
wgsl: indexing X by non-const-expr index requires X to be concrete gpuweb/gpuweb#3497, by @dneto0.
-
First pass over inline issues gpuweb/gpuweb#3496, by @kainino0x.
-
[editorial] Define "equal" for GPURenderPassLayout gpuweb/gpuweb#3495, by @kainino0x.
-
wgsl: Define trunc, ceil, floor for infinities gpuweb/gpuweb#3494, by @dneto0.
-
Describe allowance for viewport rounding gpuweb/gpuweb#3493, by @kainino0x.
-
[editorial] Work around setBindGroup overload argumentdef issue gpuweb/gpuweb#3492, by @kainino0x.
-
[editorial] Semi-formalization of interface with WGSL gpuweb/gpuweb#3491, by @kainino0x.
-
wgsl concretiation: use T2 instead of T' gpuweb/gpuweb#3490, by @dneto0.
-
Fix for two README typos gpuweb/gpuweb#3488, by @mehmetoguzderin.
-
Object labels are per-GPUObjectBase gpuweb/gpuweb#3484, by @kainino0x.
-
-
46 pull requests merged.
-
Say "indeterminate value" where previously said "undefined value" gpuweb/gpuweb#3557.
-
Add draft text/wgsl media type registration. WGSL uses UTF-8 gpuweb/gpuweb#3542.
-
Force concrete types unless all subexpressions are const-expressions gpuweb/gpuweb#3547.
-
Generalize integer texture parameter requirements gpuweb/gpuweb#3548.
-
[editorial] Clarify nan/inf eval applies to all expressions gpuweb/gpuweb#3549.
-
Fix clamp definition gpuweb/gpuweb#3553.
-
Update samples to the latest spec gpuweb/gpuweb#3551.
-
Specific GPUExternalTexture dimension in spec gpuweb/gpuweb#3544.
-
wgsl: ban ptr-to-workgroup params to user-defined functions gpuweb/gpuweb#3422.
-
wgsl: Fix how we say concrete left shift must not overflow (if known … gpuweb/gpuweb#3539.
-
wgsl: Define "indeterminate value" gpuweb/gpuweb#3508.
-
wgsl: Simplify note about runtime-sized array not being fixed-footprint gpuweb/gpuweb#3540.
-
Set GPUBuffer.[[pending_map]] to null when mapAsync() succeeds gpuweb/gpuweb#3532.
-
Rename postfix_expression to component_or_swizzle_specifier gpuweb/gpuweb#3529.
-
Remove stale/incorrect statement about inability to shadow builtins gpuweb/gpuweb#3526.
-
[editorial] Fix logo's gpuweb/gpuweb#3377.
-
workgroupStorageUsed is sum of values, each rounded up to multiple of 16 gpuweb/gpuweb#3522.
-
wgsl: bit shift width is always u32 or vector of u32 gpuweb/gpuweb#3516.
-
Fix the definition of "src-alpha-saturated" gpuweb/gpuweb#3521.
-
[editorial] Consolidate "Detailed Operations" issues into one issue gpuweb/gpuweb#3514.
-
[editorial] Misc trivial inline issue cleanup gpuweb/gpuweb#3515.
-
wgsl: Remove address space names from "other keywords" list gpuweb/gpuweb#3518.
-
Don't throw exception when getBindGroupLayout index > maxBindGroups gpuweb/gpuweb#3506.
-
Don't trail with empty BGLs in default pipeline layout gpuweb/gpuweb#3505.
-
Only allow defined GPUShaderStage bits in bind group layout entries gpuweb/gpuweb#3500.
-
Only allow defined GPUMapMode bits in mapAsync() gpuweb/gpuweb#3499.
-
[editorial] Move texture extent algorithms to texture section gpuweb/gpuweb#3512.
-
[editorial] Describe textures and subresource accounting algorithms gpuweb/gpuweb#3510.
-
[editorial] Clarify some spec around features gpuweb/gpuweb#3509.
-
Clarify behaviour of pack2x16float, quantizeToF16 for large inputs gpuweb/gpuweb#3507.
-
Require effective storage buffer binding size be a multiple of 4 gpuweb/gpuweb#3477.
-
wgsl: Define trunc, ceil, floor for infinities gpuweb/gpuweb#3494.
-
[editorial] Semi-formalization of interface with WGSL gpuweb/gpuweb#3491.
-
[editorial] Replace "super-luminant" with a more descriptive term gpuweb/gpuweb#3498.
-
[editorial] Work around setBindGroup overload argumentdef issue gpuweb/gpuweb#3492.
-
wgsl: indexing X by non-const-expr index requires X to be concrete gpuweb/gpuweb#3497.
-
[editorial] Define "equal" for GPURenderPassLayout gpuweb/gpuweb#3495.
-
First pass over inline issues gpuweb/gpuweb#3496.
-
Rename maxColorAttachmentBytesPerPixel to *PerSample gpuweb/gpuweb#3434.
-
Matching rules for fragment shader output and color attachments gpuweb/gpuweb#3408.
-
Object labels are per-GPUObjectBase gpuweb/gpuweb#3484.
-
Describe allowance for viewport rounding gpuweb/gpuweb#3493.
-
Rewwrite atan2 definition gpuweb/gpuweb#3474.
-
wgsl concretiation: use T2 instead of T' gpuweb/gpuweb#3490.
-
Fix for two README typos gpuweb/gpuweb#3488.
-
Clarify definition of the concretization of a type. gpuweb/gpuweb#3482.
-
-
1 pull request closed.
-
Streamline tokens gpuweb/gpuweb#3480.
-
-
1 pull request reopened.
-
Streamline tokens gpuweb/gpuweb#3480, by @mehmetoguzderin.
-
-
30 pull requests received 80 new comments.
-
wgsl: Define "indeterminate value" gpuweb/gpuweb#3508 (15 comments by @alan-baker, @dneto0, @jimblandy, @kdashg)
-
Add draft text/wgsl media type registration. WGSL uses UTF-8 gpuweb/gpuweb#3542 (11 comments by @annevk, @dneto0, @mehmetoguzderin)
-
WGSL syntax styling upgrade with initial notation section gpuweb/gpuweb#3543 (4 comments by @alan-baker, @dneto0, @mehmetoguzderin)
-
Generalize integer texture parameter requirements gpuweb/gpuweb#3548 (4 comments by @ben-clayton, @dneto0)
-
wgsl: bit shift width is always u32 or vector of u32 gpuweb/gpuweb#3516 (3 comments by @dneto0, @kdashg)
-
Remove special-case uniformity rules for short-circuiting ops gpuweb/gpuweb#3558 (2 comments by @dneto0, @jrprice)
-
[editorial] Clarify nan/inf eval applies to all expressions gpuweb/gpuweb#3549 (2 comments by @alan-baker, @dneto0)
-
wgsl: Fix how we say concrete left shift must not overflow (if known … gpuweb/gpuweb#3539 (2 comments by @dneto0)
-
workgroupStorageUsed is sum of values, each rounded up to multiple of 16 gpuweb/gpuweb#3522 (2 comments by @kainino0x, @kdashg)
-
Only allow defined GPUShaderStage bits in bind group layout entries gpuweb/gpuweb#3500 (2 comments by @kainino0x)
-
Require effective storage buffer binding size be a multiple of 4 gpuweb/gpuweb#3477 (2 comments by @Kangz, @kdashg)
-
Clarify definition of the concretization of a type. gpuweb/gpuweb#3482 (2 comments by @dj2, @dneto0)
-
Force concrete types unless all subexpressions are const-expressions gpuweb/gpuweb#3547 (1 comment by @ben-clayton)
-
Remove stale/incorrect statement about inability to shadow builtins gpuweb/gpuweb#3526 (1 comment by @mehmetoguzderin)
-
Don't throw exception when getBindGroupLayout index > maxBindGroups gpuweb/gpuweb#3506 (1 comment by @kainino0x)
-
gpuweb/cts
-
18 issues opened.
-
wgsl: wgsl source must e UTF-8 without a byte order mark gpuweb/cts#1949, by @dneto0.
-
quantizeToF16 tests do not correctly handle FTZ for f16 denorms gpuweb/cts#1946, by @zoddicus.
-
GPUBuffer.unmap() shouldn't fail even on unmapped or destroyed buffers gpuweb/cts#1930, by @takahirox.
-
Debug marker validation and operation tests gpuweb/cts#1928, by @Gyuyoung.
-
val: pipeline_bind_group_compat gpuweb/cts#1916, by @kainino0x.
-
val: encoding,encoder_state gpuweb/cts#1914, by @kainino0x.
-
wgsl: (1 << 2u) is abstract value 4: RHS of shifts must be u32 or vector of u32 gpuweb/cts#1909, by @dneto0.
-
Add validation tests on effective buffer binding size gpuweb/cts#1905, by @Jiawei-Shao.
-
op: Regression test for Adreno depth synchronization bug gpuweb/cts#1904, by @kainino0x.
-
val: setBindGroup required for a bind group layout that exists but is empty gpuweb/cts#1903, by @kainino0x.
-
wgsl: address space names are not keywords gpuweb/cts#1902, by @dneto0.
-
wgsl: indexing X by non-const-expr index requires X to be concrete gpuweb/cts#1893, by @dneto0.
-
Check the validation of getBindGroupLayout method gpuweb/cts#1891, by @Gyuyoung.
-
wgsl: execution test for pack2x16float builtin gpuweb/cts#1889, by @zoddicus.
-
-
12 issues closed.
-
wgsl: execution test for unpack2x16snorm gpuweb/cts#1291.
-
wgsl: execution test for unpack4x8unorm builtin gpuweb/cts#1290.
-
wgsl: execution test for unpack2x16unorm builtin gpuweb/cts#1292.
-
wgsl: execution test for pack4x8unorm builtin gpuweb/cts#1285.
-
wgsl: execution test for pack4x8snorm builtin gpuweb/cts#1284.
-
wgsl: execution test for pack2x16snorm gpuweb/cts#1286.
-
wgsl: execution test for pack2x16unorm builtin gpuweb/cts#1287.
-
Add validation tests on effective buffer binding size gpuweb/cts#1905.
-
wgsl: execution test for pack2x16float builtin gpuweb/cts#1889.
-
Extend the ShaderModuleEntryPoint tests with Unicode gpuweb/cts#1130.
-
val: writeTexture validation that isn't common with copyB2T/T2B gpuweb/cts#1692.
-
17 issues received 24 new comments.
-
Add validation tests on effective buffer binding size gpuweb/cts#1905 (2 comments by @Jiawei-Shao, @kainino0x)
-
val: setBindGroup required for a bind group layout that exists but is empty gpuweb/cts#1903 (2 comments by @Gyuyoung, @kainino0x)
-
op: pipeline,default_layout:layout - default layout is computed correctly gpuweb/cts#924 (1 comment by @lokokung)
-
op: render_pipeline and compute_pipeline entry_point_name tests gpuweb/cts#895 (1 comment by @lokokung)
-
val: writeTexture validation that isn't common with copyB2T/T2B gpuweb/cts#1692 (1 comment by @Gyuyoung)
-
45 pull requests opened.
-
Implement `unpack4x8snorm` tests gpuweb/cts#1950, by @zoddicus.
-
val: Add a control case to 'pass_end_invalid_order' test gpuweb/cts#1948, by @Gyuyoung.
-
Properly handle potential FTZ in quantizeToF16 tests gpuweb/cts#1947, by @zoddicus.
-
op: Introduce debug_marker.spec.ts to test operations of debug APIs gpuweb/cts#1945, by @Gyuyoung.
-
Handle error from division in unpack2x16unorm gpuweb/cts#1944, by @zoddicus.
-
Implement `unpack4x8unorm` tests gpuweb/cts#1943, by @zoddicus.
-
Change from using `full` to `dense` for utility function names gpuweb/cts#1942, by @zoddicus.
-
Implement `unpack2x16unorm` tests gpuweb/cts#1941, by @zoddicus.
-
Fixup min i32 value in parsing test. gpuweb/cts#1940, by @dj2.
-
wgsl: Add robustness tests for builtin shadowing gpuweb/cts#1939, by @ben-clayton.
-
Add rg11b10ufloat-renderable validation test plan gpuweb/cts#1937, by @takahirox.
-
Use cast instead of suffix for signed integers in Scalar gpuweb/cts#1936, by @zoddicus.
-
Implement `unpack2x16snorm` tests gpuweb/cts#1935, by @zoddicus.
-
Implement `unpack2x16float` tests gpuweb/cts#1934, by @zoddicus.
-
Convert f32 binary expression tests to use pre-generated inputs gpuweb/cts#1933, by @zoddicus.
-
Convert pack* tests to use generated vectors for inputs gpuweb/cts#1932, by @zoddicus.
-
val: Introduce debugMarker.spec.ts to test debug marker gpuweb/cts#1931, by @Gyuyoung.
-
GPUBuffer.unmap() shouldn't fail even on unmapped or destroyed buffers gpuweb/cts#1929, by @takahirox.
-
Implement `pack4x8snorm` tests gpuweb/cts#1924, by @zoddicus.
-
Implement `pack4x8unorm` tests gpuweb/cts#1923, by @zoddicus.
-
Implement `pack2x16snorm` tests gpuweb/cts#1922, by @zoddicus.
-
val: Introduce encoder_open_state.spec.ts to test when an encoder is open gpuweb/cts#1921, by @Gyuyoung.
-
Implement `pack2x16unorm` tests gpuweb/cts#1920, by @zoddicus.
-
val: Implement 'call_after_successful_finish' test gpuweb/cts#1918, by @Gyuyoung.
-
val: Implement 'pass_end_none' and 'pass_end_twice' tests gpuweb/cts#1917, by @Gyuyoung.
-
val: Implement 'begin_render_pass_before_ending_previous_pass' test gpuweb/cts#1913, by @Gyuyoung.
-
val: Implement 'pass_end_invalid_order' test gpuweb/cts#1912, by @Gyuyoung.
-
Fix pack2x16float tests to correctly handle corner cases gpuweb/cts#1911, by @zoddicus.
-
val: Implement 'empty_bind_group_layouts_requires_empty_bind_groups' test gpuweb/cts#1910, by @Gyuyoung.
-
Implement `reflect` f32 tests gpuweb/cts#1908, by @zoddicus.
-
Add tests on effective buffer binding size the fix existing tests gpuweb/cts#1906, by @Jiawei-Shao.
-
Implement `faceForward` f32 tests gpuweb/cts#1900, by @zoddicus.
-
Implement 'index_range,auto_layout' test in getBindGroupLayout.spec.ts gpuweb/cts#1899, by @Gyuyoung.
-
Implement `asin` f32 tests gpuweb/cts#1898, by @zoddicus.
-
Implement `acos` f32 tests gpuweb/cts#1897, by @zoddicus.
-
Remove a fallback value in the use of maxBindGroups value gpuweb/cts#1895, by @Gyuyoung.
-
Implement `distance` f32 tests gpuweb/cts#1894, by @zoddicus.
-
Implement 'index_range' test in getBindGroupLayout.spec.ts gpuweb/cts#1892, by @Gyuyoung.
-
Implement `pack2x16float` tests gpuweb/cts#1890, by @zoddicus.
-
Add rg11b10ufloat-renderable feature requestDevice test gpuweb/cts#1888, by @takahirox.
-
-
40 pull requests merged.
-
val: Add a control case to 'pass_end_invalid_order' test gpuweb/cts#1948.
-
wgsl: Add robustness tests for builtin shadowing gpuweb/cts#1939.
-
Implement `unpack2x16snorm` tests gpuweb/cts#1935.
-
Handle error from division in unpack2x16unorm gpuweb/cts#1944.
-
Implement `unpack4x8unorm` tests gpuweb/cts#1943.
-
Implement `unpack2x16unorm` tests gpuweb/cts#1941.
-
Fixup min i32 value in parsing test. gpuweb/cts#1940.
-
val: Introduce debugMarker.spec.ts to test debug marker gpuweb/cts#1931.
-
Use cast instead of suffix for signed integers in Scalar gpuweb/cts#1936.
-
Convert f32 binary expression tests to use pre-generated inputs gpuweb/cts#1933.
-
Convert pack* tests to use generated vectors for inputs gpuweb/cts#1932.
-
Implement `pack4x8unorm` tests gpuweb/cts#1923.
-
Implement `pack4x8snorm` tests gpuweb/cts#1924.
-
Implement `pack2x16snorm` tests gpuweb/cts#1922.
-
Implement `pack2x16unorm` tests gpuweb/cts#1920.
-
Implement `faceForward` f32 tests gpuweb/cts#1900.
-
val: Implement 'call_after_successful_finish' test gpuweb/cts#1918.
-
val: Implement 'pass_end_invalid_order' test gpuweb/cts#1912.
-
Implement `reflect` f32 tests gpuweb/cts#1908.
-
Implement `distance` f32 tests gpuweb/cts#1894.
-
Fix pack2x16float tests to correctly handle corner cases gpuweb/cts#1911.
-
Implement `asin` f32 tests gpuweb/cts#1898.
-
Implement `acos` f32 tests gpuweb/cts#1897.
-
Add tests on effective buffer binding size the fix existing tests gpuweb/cts#1906.
-
Implement `pack2x16float` tests gpuweb/cts#1890.
-
Implement 'index_range' test in getBindGroupLayout.spec.ts gpuweb/cts#1892.
-
Add rg11b10ufloat-renderable feature requestDevice test gpuweb/cts#1888.
-
op: Implement 'ww' test in multiple_buffers.spec.ts gpuweb/cts#1879.
-
Remove a fallback value in the use of maxBindGroups value gpuweb/cts#1895.
-
Update `atan2` F32 tests to account for atan based implementations gpuweb/cts#1837.
-
Implement `cross` f32 tests gpuweb/cts#1878.
-
compilation_info: Check offset when there is a carriage-return. gpuweb/cts#1880.
-
-
3 pull requests closed.
-
op: Introduce debug_marker.spec.ts to test operations of debug APIs gpuweb/cts#1945.
-
Change from using `full` to `dense` for utility function names gpuweb/cts#1942.
-
val: Implement 'begin_render_pass_before_ending_previous_pass' test gpuweb/cts#1913.
-
-
32 pull requests received 161 new comments.
-
val: Introduce encoder_open_state.spec.ts to test when an encoder is open gpuweb/cts#1921 (22 comments by @Gyuyoung, @kainino0x)
-
Fix pack2x16float tests to correctly handle corner cases gpuweb/cts#1911 (14 comments by @kainino0x, @zoddicus)
-
Implement 'index_range' test in getBindGroupLayout.spec.ts gpuweb/cts#1892 (14 comments by @Gyuyoung, @kainino0x)
-
val: Implement 'empty_bind_group_layouts_requires_empty_bind_groups,render_pass' test gpuweb/cts#1915 (9 comments by @Gyuyoung, @kainino0x)
-
op: Implement 'two_dispatches_in_the_same_compute_pass' test in multiple_buffers.spec.ts gpuweb/cts#1896 (5 comments by @Gyuyoung, @kainino0x)
-
op: Introduce debug_marker.spec.ts to test operations of debug APIs gpuweb/cts#1945 (4 comments by @Gyuyoung, @kainino0x)
-
val: Implement 'pass_end_none' and 'pass_end_twice' tests gpuweb/cts#1917 (4 comments by @Gyuyoung, @kainino0x)
-
Change from using `full` to `dense` for utility function names gpuweb/cts#1942 (4 comments by @ben-clayton, @zoddicus)
-
val: Use doCompute,doRender helpers in empty_bind_group_layouts_requires_empty_bind_groups.* tests gpuweb/cts#1919 (4 comments by @Gyuyoung, @kainino0x)
-
val: Implement 'empty_bind_group_layouts_requires_empty_bind_groups' test gpuweb/cts#1910 (4 comments by @Gyuyoung, @kainino0x)
-
Implement 'index_range,auto_layout' test in getBindGroupLayout.spec.ts gpuweb/cts#1899 (4 comments by @Gyuyoung, @kainino0x)
-
val: Introduce debugMarker.spec.ts to test debug marker gpuweb/cts#1931 (2 comments by @Gyuyoung, @toji)
-
val: Implement 'call_after_successful_finish' test gpuweb/cts#1918 (2 comments by @Gyuyoung, @kainino0x)
-
Add tests on effective buffer binding size the fix existing tests gpuweb/cts#1906 (2 comments by @jiangzhaoming, @Jiawei-Shao)
-
Update `atan2` F32 tests to account for atan based implementations gpuweb/cts#1837 (2 comments by @dneto0, @zoddicus)
-
Lifecycle
-
1 charter extension.
-
Charter extended until 30 November 2022.
-
TR publications
-
15 documents published.
-
WebGPU published 4 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 11 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/cts
-
21 issues opened.
-
op: adapter becomes stale when device is destroyed gpuweb/cts#1886, by @kainino0x.
-
web_platform: canvas,getPreferredFormat gpuweb/cts#1885, by @kainino0x.
-
web_platform: canvas,configure gpuweb/cts#1884, by @kainino0x.
-
val: initialization,requestDevice:limits gpuweb/cts#1883, by @kainino0x.
-
val: initialization,requestDevice:features,nonexistent gpuweb/cts#1882, by @kainino0x.
-
val: initialization,requestDevice:features,known_but_unavailable gpuweb/cts#1881, by @kainino0x.
-
wgsl: execution tests for boolean equality/inequality gpuweb/cts#1875, by @zoddicus.
-
Add rg11b10ufloat-renderable feature test gpuweb/cts#1874, by @takahirox.
-
Add rg11b10ufloat-renderable feature tests gpuweb/cts#1873, by @takahirox.
-
wgsl: execution tests for Logical Expressions gpuweb/cts#1868, by @zoddicus.
-
creation-time bitcast to a NaN results in a shader-creation error gpuweb/cts#1859, by @dneto0.
-
wgsl: Test for array arguments being passed by copy gpuweb/cts#1855, by @ben-clayton.
-
wgsl: if LHS is signed, the right shift is arithmetic gpuweb/cts#1853, by @dneto0.
-
wgsl: user-defined functions with ptr params must satisfy aliasing restrictions gpuweb/cts#1852, by @dneto0.
-
Tracking issue for overrides NaN/Inf value validation gpuweb/cts#1838, by @shrekshao.
-
op: rendering,blending gpuweb/cts#1835, by @kainino0x.
-
30 issues closed.
-
web_platform: reference tests imported to WPT need to be fuzzy gpuweb/cts#1045.
-
reftests: test more canvas texture formats gpuweb/cts#915.
-
reftests: test canvas compositingAlphaMode gpuweb/cts#917.
-
val: cmds,buffer_texture_copies should move to image_copy gpuweb/cts#1009.
-
op: buffer/device.destroy() unmaps buffer on current thread gpuweb/cts#1697.
-
op: requestAdapter gpuweb/cts#1061.
-
op: explicitly test WebGPU is available on DedicatedWorkers gpuweb/cts#1066.
-
val: Test beginComputePass and GPUComputePassDescriptor validation gpuweb/cts#1743.
-
Update val:CopyExternalImageToTexture:OOB,source gpuweb/cts#1738.
-
val: Test copyExternalImageToTexture with dimensions gpuweb/cts#1694.
-
val: render pass descriptor gpuweb/cts#1618.
-
op: adapter becomes stale when device is destroyed gpuweb/cts#1886.
-
web_platform: canvas,getPreferredFormat gpuweb/cts#1885.
-
web_platform: canvas,configure gpuweb/cts#1884.
-
val: initialization,requestDevice:limits gpuweb/cts#1883.
-
val: initialization,requestDevice:features,nonexistent gpuweb/cts#1882.
-
val: initialization,requestDevice:features,known_but_unavailable gpuweb/cts#1881.
-
val: createBindGroup gpuweb/cts#884.
-
val: createBindGroupLayout gpuweb/cts#885.
-
wgsl: execution tests for boolean equality/inequality gpuweb/cts#1875.
-
val: GPURenderPassDescriptor.maxDrawCount test gpuweb/cts#1613.
-
Add rg11b10ufloat-renderable feature tests gpuweb/cts#1873.
-
wgsl: execution tests for Logical Expressions gpuweb/cts#1868.
-
Tracking issue for overrides NaN/Inf value validation gpuweb/cts#1838.
-
wgsl: Test for array arguments being passed by copy gpuweb/cts#1855.
-
wgsl: execution test quantizeToF16 gpuweb/cts#1200.
-
wgsl: Add idiomatic support for f16 values gpuweb/cts#1812.
-
-
1 issue reopened.
-
val: device_mismatch throughout CTS gpuweb/cts#912, by @kainino0x.
-
-
39 issues received 48 new comments.
-
webgpu:api,validation,queue,writeBuffer:ranges: tries to `writeBuffer` with Float16Array gpuweb/cts#1832 (4 comments by @austinEng, @kainino0x, @zoddicus)
-
canvas read as an image source alpha mode behavior update gpuweb/cts#1561 (2 comments by @kainino0x, @shrekshao)
-
val: writeTexture validation that isn't common with copyB2T/T2B gpuweb/cts#1692 (1 comment by @kainino0x)
-
web_platform: reference tests imported to WPT need to be fuzzy gpuweb/cts#1045 (1 comment by @kainino0x)
-
op: render_pipeline and compute_pipeline entry_point_name tests gpuweb/cts#895 (1 comment by @kainino0x)
-
val: Test beginComputePass and GPUComputePassDescriptor validation gpuweb/cts#1743 (1 comment by @kainino0x)
-
op: device.lost returns the same object each time (before or after destruction) gpuweb/cts#1887 (1 comment by @kainino0x)
-
val: initialization,requestDevice:features,known_but_unavailable gpuweb/cts#1881 (1 comment by @kainino0x)
-
Test bug: bind group layout with sampleType: "float" and multisampled: true is invalid gpuweb/cts#1848 (1 comment by @kainino0x)
-
op: Allow bit patterns for the same value in copyTextureToTexture gpuweb/cts#973 (1 comment by @kainino0x)
-
-
49 pull requests opened.
-
compilation_info: Check offset when there is a carriage-return. gpuweb/cts#1880, by @Kangz.
-
op: Implement 'ww' test in multiple_buffers.spec.ts gpuweb/cts#1879, by @Gyuyoung.
-
Implement `cross` f32 tests gpuweb/cts#1878, by @zoddicus.
-
Implement 'timestampWrites,invalid_query_set' test in beginComputePass.spec.ts gpuweb/cts#1877, by @Gyuyoung.
-
Implement boolean logical equality tests gpuweb/cts#1876, by @zoddicus.
-
op: Implement 'wr' test in multiple_buffers.spec.ts gpuweb/cts#1872, by @Gyuyoung.
-
Style and bug fixes in device_mismatch tests gpuweb/cts#1871, by @kainino0x.
-
Add execution tests for logical binary operations gpuweb/cts#1870, by @zoddicus.
-
Add execution tests for logical negation gpuweb/cts#1869, by @zoddicus.
-
Add a 'sampler,device_mismatch' test to createBindGroup.spec.ts gpuweb/cts#1867, by @Gyuyoung.
-
Update align CTS to match spec change for u32 values. gpuweb/cts#1866, by @dj2.
-
Remove copyToTexture,CopyExternalImageToTexture:destination_texture,d… gpuweb/cts#1864, by @Jiawei-Shao.
-
wgsl: Remove gradients_in_varying_loop tests gpuweb/cts#1863, by @ben-clayton.
-
wgsl: Fix emission of f32 and f16 infinities. gpuweb/cts#1862, by @ben-clayton.
-
webgpu/shader/execution: Reduce `const` case batch size gpuweb/cts#1860, by @ben-clayton.
-
op: Implement test in multiple_buffers.spec.ts gpuweb/cts#1857, by @Gyuyoung.
-
Update canvas_composite_alpha to test opaque alpha mode canvas composite gpuweb/cts#1856, by @shrekshao.
-
Update align tests for const expressions. gpuweb/cts#1854, by @dj2.
-
Update multisampled_validation test to check the sampleType as well gpuweb/cts#1850, by @Gyuyoung.
-
Add the buffer_dynamic_offsets test to setBindGroup.spec.ts gpuweb/cts#1849, by @Gyuyoung.
-
Add a doc about Floating Point math as background on numeric testing gpuweb/cts#1847, by @zoddicus.
-
Implement test in operation/rendering/blending.spec.ts gpuweb/cts#1845, by @Gyuyoung.
-
fix dev server for float16.js gpuweb/cts#1843, by @kainino0x.
-
Overrides: add operation tests for compute pipeline gpuweb/cts#1842, by @shrekshao.
-
Implement `normalize` f32 tests gpuweb/cts#1841, by @zoddicus.
-
Fix DeviceLostTests if `noRaceWithRejectOnTimeout == true` gpuweb/cts#1840, by @austinEng.
-
Add RENDER_ATTACHMENT usage in createBindGroup:multisampled_validation:* gpuweb/cts#1839, by @Jiawei-Shao.
-
Update `atan2` F32 tests to account for atan based implementations gpuweb/cts#1837, by @zoddicus.
-
Add global option noRaceWithRejectOnTimeout gpuweb/cts#1836, by @austinEng.
-
shader/execution: Add heartbeats to expression tests gpuweb/cts#1834, by @ben-clayton.
-
Fix type error when trying to .writeBuffer a Float16Array gpuweb/cts#1833, by @zoddicus.
-
Fix f16 hex literal testcases of having suffix "h" without "p" gpuweb/cts#1831, by @jiangzhaoming.
-
Add pipeline overridable constants validation tests gpuweb/cts#1830, by @shrekshao.
-
Add a 'storage_texture,format' test to createBindGroup.spec.ts gpuweb/cts#1829, by @Gyuyoung.
-
Implement `quantizeToF16` f32 tests gpuweb/cts#1828, by @zoddicus.
-
shader/execution: Fix shader creation errors gpuweb/cts#1827, by @ben-clayton.
-
Add 'buffer,resource_binding_size' to createBindGroup.spec.ts gpuweb/cts#1826, by @Gyuyoung.
-
Add 'buffer,resource_offset' to createBindGroup.spec.ts gpuweb/cts#1825, by @Gyuyoung.
-
Update 'buffer,usage' test in createBindGroup.spec.ts gpuweb/cts#1824, by @Gyuyoung.
-
slightly broaden/simplify pipeline_output_targets gpuweb/cts#1823, by @kainino0x.
-
test ~all cases of pipeline_output_targets,blend gpuweb/cts#1822, by @kainino0x.
-
flake repro: 208 subcases gpuweb/cts#1821, by @kainino0x.
-
Add 'buffer,usage' test to createBindGroup.spec.ts gpuweb/cts#1820, by @Gyuyoung.
-
Add storage_texture,mip_level_count test to createBindGroup.spec.ts gpuweb/cts#1818, by @Gyuyoung.
-
-
48 pull requests merged.
-
op: Implement 'wr' test in multiple_buffers.spec.ts gpuweb/cts#1872.
-
Implement boolean logical equality tests gpuweb/cts#1876.
-
Style and bug fixes in device_mismatch tests gpuweb/cts#1871.
-
Add a 'sampler,device_mismatch' test to createBindGroup.spec.ts gpuweb/cts#1867.
-
op: Implement test in multiple_buffers.spec.ts gpuweb/cts#1857.
-
Add execution tests for logical binary operations gpuweb/cts#1870.
-
Add execution tests for logical negation gpuweb/cts#1869.
-
Update align CTS to match spec change for u32 values. gpuweb/cts#1866.
-
wgsl: Remove gradients_in_varying_loop tests gpuweb/cts#1863.
-
webgpu/shader/execution: Reduce `const` case batch size gpuweb/cts#1860.
-
wgsl: Fix emission of f32 and f16 infinities. gpuweb/cts#1862.
-
Add a doc about Floating Point math as background on numeric testing gpuweb/cts#1847.
-
Update align tests for const expressions. gpuweb/cts#1854.
-
Implement test in operation/rendering/blending.spec.ts gpuweb/cts#1845.
-
Update multisampled_validation test to check the sampleType as well gpuweb/cts#1850.
-
Overrides: add operation tests for compute pipeline gpuweb/cts#1842.
-
Add the buffer_dynamic_offsets test to setBindGroup.spec.ts gpuweb/cts#1849.
-
Add pipeline overridable constants validation tests gpuweb/cts#1830.
-
Implement `normalize` f32 tests gpuweb/cts#1841.
-
fix dev server for float16.js gpuweb/cts#1843.
-
Add 'buffer,resource_binding_size' to createBindGroup.spec.ts gpuweb/cts#1826.
-
Fix f16 hex literal testcases of having suffix "h" without "p" gpuweb/cts#1831.
-
Add global option noRaceWithRejectOnTimeout gpuweb/cts#1836.
-
Add a 'storage_texture,format' test to createBindGroup.spec.ts gpuweb/cts#1829.
-
Fix type error when trying to .writeBuffer a Float16Array gpuweb/cts#1833.
-
shader/execution: Add heartbeats to expression tests gpuweb/cts#1834.
-
Add `length` F32 tests gpuweb/cts#1815.
-
Make `make*ToF32IntervalCase`s more idiomatic gpuweb/cts#1816.
-
Implement `quantizeToF16` f32 tests gpuweb/cts#1828.
-
Update 'buffer,usage' test in createBindGroup.spec.ts gpuweb/cts#1824.
-
Add 'buffer,resource_offset' to createBindGroup.spec.ts gpuweb/cts#1825.
-
shader/execution: Fix shader creation errors gpuweb/cts#1827.
-
Add 'buffer,usage' test to createBindGroup.spec.ts gpuweb/cts#1820.
-
slightly broaden/simplify pipeline_output_targets gpuweb/cts#1823.
-
test ~all cases of pipeline_output_targets,blend gpuweb/cts#1822.
-
Add storage_texture,mip_level_count test to createBindGroup.spec.ts gpuweb/cts#1818.
-
Minor clean up in life_of.md gpuweb/cts#1608.
-
Add storage_texture,usage test to createVindGroup.spec.ts gpuweb/cts#1813.
-
1 pull request closed.
-
Fix DeviceLostTests if `noRaceWithRejectOnTimeout == true` gpuweb/cts#1840.
-
-
32 pull requests received 223 new comments.
-
Add a doc about Floating Point math as background on numeric testing gpuweb/cts#1847 (88 comments by @alan-baker, @ben-clayton, @dneto0, @zoddicus)
-
Add the buffer_dynamic_offsets test to setBindGroup.spec.ts gpuweb/cts#1849 (13 comments by @Gyuyoung, @kainino0x)
-
Update `atan2` F32 tests to account for atan based implementations gpuweb/cts#1837 (11 comments by @dneto0, @zoddicus)
-
Add pipeline overridable constants validation tests gpuweb/cts#1830 (9 comments by @kainino0x, @shrekshao)
-
Implement test in operation/rendering/blending.spec.ts gpuweb/cts#1845 (8 comments by @Gyuyoung, @kainino0x)
-
op: Implement test in multiple_buffers.spec.ts gpuweb/cts#1857 (6 comments by @Gyuyoung, @kainino0x)
-
Overrides: add operation tests for compute pipeline gpuweb/cts#1842 (6 comments by @kainino0x, @shrekshao)
-
Add 'buffer,resource_binding_size' to createBindGroup.spec.ts gpuweb/cts#1826 (6 comments by @Gyuyoung, @kainino0x)
-
Update 'buffer,usage' test in createBindGroup.spec.ts gpuweb/cts#1824 (6 comments by @Gyuyoung, @kainino0x)
-
Implement 'timestampWrites,invalid_query_set' test in beginComputePass.spec.ts gpuweb/cts#1877 (5 comments by @Gyuyoung, @kainino0x)
-
Update 'multisampled_validation' test in createBindGroupLayout.spec.ts gpuweb/cts#1777 (4 comments by @Gyuyoung, @kainino0x)
-
Update multisampled_validation test to check the sampleType as well gpuweb/cts#1850 (4 comments by @Gyuyoung, @kainino0x)
-
webgpu/shader/execution: Reduce `const` case batch size gpuweb/cts#1860 (3 comments by @austinEng, @ben-clayton)
-
Add 'buffer,usage' test to createBindGroup.spec.ts gpuweb/cts#1820 (3 comments by @Gyuyoung, @kainino0x)
-
Fix a wrong bit operation to decide the validation in indirect_dispatch_buffer,usage gpuweb/cts#1819 (3 comments by @Gyuyoung, @kainino0x)
-
Add storage_texture,mip_level_count test to createBindGroup.spec.ts gpuweb/cts#1818 (3 comments by @Gyuyoung, @kainino0x)
-
Update texture_must_have_correct_dimension test to handle storageTexture as well gpuweb/cts#1851 (2 comments by @Gyuyoung, @kainino0x)
-
Fix DeviceLostTests if `noRaceWithRejectOnTimeout == true` gpuweb/cts#1840 (2 comments by @kainino0x)
-
Fix type error when trying to .writeBuffer a Float16Array gpuweb/cts#1833 (2 comments by @kainino0x, @zoddicus)
-
Add a 'storage_texture,format' test to createBindGroup.spec.ts gpuweb/cts#1829 (2 comments by @Gyuyoung, @kainino0x)
-
Add 'buffer,resource_offset' to createBindGroup.spec.ts gpuweb/cts#1825 (2 comments by @Gyuyoung, @kainino0x)
-
compilation_info: Check offset when there is a carriage-return. gpuweb/cts#1880 (1 comment by @Kangz)
-
Add RENDER_ATTACHMENT usage in createBindGroup:multisampled_validation:* gpuweb/cts#1839 (1 comment by @Gyuyoung)
-
Fix f16 hex literal testcases of having suffix "h" without "p" gpuweb/cts#1831 (1 comment by @jiangzhaoming)
-
gpuweb/gpuweb
-
23 issues opened.
-
Uniformity analysis is far too strict gpuweb/gpuweb#3479, by @brendan-duncan.
-
Enabling conservative depth gpuweb/gpuweb#3476, by @jonpry.
-
why can't *any* fixed-size array use an element count that is an override expression gpuweb/gpuweb#3475, by @dneto0.
-
Definition of "clamp-to-edge" is not very precise gpuweb/gpuweb#3473, by @kainino0x.
-
`textureSampleBaseClampToEdge` clamping should clamp a half-texel from edge gpuweb/gpuweb#3471, by @ben-clayton.
-
Clarify intended behaviour when packing out of range values gpuweb/gpuweb#3470, by @zoddicus.
-
Bitcasting to NaN should be a compile error? gpuweb/gpuweb#3468, by @litherum.
-
Allow attributes to be abstract, i32 or u32. gpuweb/gpuweb#3465, by @dj2.
-
wgsl: somehow we lost arithmetic right shift gpuweb/gpuweb#3461, by @dneto0.
-
wgsl editorial: uniformity rules say "reference to ...." and that's confusing gpuweb/gpuweb#3457, by @dneto0.
-
Does getBindGroupLayout return the same underlying object or a new one? gpuweb/gpuweb#3454, by @kainino0x.
-
Inconsistencies in createBuffer validation gpuweb/gpuweb#3453, by @kainino0x.
-
wgsl: clarify array type equivalence when element count is overridable gpuweb/gpuweb#3441, by @dneto0.
-
wgsl: atan2 result depends on signs of both arguments, not spec'd gpuweb/gpuweb#3437, by @dneto0.
-
Override constant should not provided by pipeline as NaN or Inf gpuweb/gpuweb#3435, by @jiangzhaoming.
-
Difference between `src-alpha` and `src-alpha-saturated` gpuweb/gpuweb#3432, by @dj2.
-
wgsl: add way to get special values for a given numeric type (e.g. max, min, lowest) gpuweb/gpuweb#3431, by @dneto0.
-
GPURenderBundleEncoder.setPipeline() not implementable on MTLIndirectCommandBuffer gpuweb/gpuweb#3423, by @litherum.
-
Clarification needed with regards to flushing denorm f16 values gpuweb/gpuweb#3421, by @zoddicus.
-
WGSL spec gives readers the runaround regarding `var` access modes gpuweb/gpuweb#3419, by @jimblandy.
-
Text still assumes read-only storage textures are allowed gpuweb/gpuweb#3418, by @jimblandy.
-
Inconsistent status of leading zeros gpuweb/gpuweb#3415, by @mehmetoguzderin.
-
Definition of length(x) is too restrictive for x being a scalar gpuweb/gpuweb#3407, by @zoddicus.
-
-
41 issues closed.
-
Override constant should not provided by pipeline as NaN or Inf gpuweb/gpuweb#3435.
-
Allow attributes to be abstract, i32 or u32. gpuweb/gpuweb#3465.
-
`textureSampleBaseClampToEdge` clamping should clamp a half-texel from edge gpuweb/gpuweb#3471.
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766.
-
Is an atomic a composite type? gpuweb/gpuweb#3004.
-
Conversion expressions for the Abstract types gpuweb/gpuweb#2959.
-
Missing definition of `T` in accessor expressions gpuweb/gpuweb#3208.
-
Constructor expressions use `T` without defining `T` gpuweb/gpuweb#3127.
-
Is `@align(1)` valid? gpuweb/gpuweb#3131.
-
Clarify storable type concrete components for vectors and matrices. gpuweb/gpuweb#3031.
-
wgsl: Atomic Types should mention initialization gpuweb/gpuweb#2977.
-
Add text and example around IO locations spanning multiple locations gpuweb/gpuweb#3045.
-
`ldexp` needs clarification gpuweb/gpuweb#3134.
-
Add texture types to `t` for textureSampleGrad and textureSampleBias gpuweb/gpuweb#2892.
-
Clarification of note in fixed-footprint types section gpuweb/gpuweb#3024.
-
Clarify use of variable in memory overlap and pointers sections gpuweb/gpuweb#3026.
-
Clarify meaning of `context` in integer example in `TYPE INFERENCE FOR LITERALS` gpuweb/gpuweb#2970.
-
Clarify note about array type with overrideable constant sizing gpuweb/gpuweb#3023.
-
Update note about when automatic conversions occur. gpuweb/gpuweb#2939.
-
wgsl: somehow we lost arithmetic right shift gpuweb/gpuweb#3461.
-
Clarification needed with regards to flushing denorm f16 values gpuweb/gpuweb#3421.
-
Clarify the normalized float ranges in the `pack` methods gpuweb/gpuweb#2903.
-
Grammar update in for statement gpuweb/gpuweb#3186.
-
`external` textures and `textureSample` gpuweb/gpuweb#2894.
-
editorial bugs in WGSL spec gpuweb/gpuweb#2920.
-
wgsl: rename type_decl grammar element to type_specifier gpuweb/gpuweb#3356.
-
wgsl: clarify array type equivalence when element count is overridable gpuweb/gpuweb#3441.
-
Change ArrayBuffer arguments to ArrayBuffer + srcOffset gpuweb/gpuweb#244.
-
Describe more clearly who can contribute in README.md gpuweb/gpuweb#2237.
-
Colorspace conversions for GPUExternalTexture are arbitrarily complex gpuweb/gpuweb#3293.
-
sampleType: "float" is not valid with multisampled: true gpuweb/gpuweb#3364.
-
Overloading of the word `pipeline` gpuweb/gpuweb#3034.
-
Clarification on `const` in static access gpuweb/gpuweb#3204.
-
wgsl: compute shaders have no user-defined inputs or outputs gpuweb/gpuweb#3227.
-
WGSL spec gives readers the runaround regarding `var` access modes gpuweb/gpuweb#3419.
-
Definition of length(x) is too restrictive for x being a scalar gpuweb/gpuweb#3407.
-
Inconsistent status of leading zeros gpuweb/gpuweb#3415.
-
[editorial] unmap should check for device loss gpuweb/gpuweb#3402.
-
-
5 issues reopened.
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766, by @dneto0.
-
Override constant should not provided by pipeline as NaN or Inf gpuweb/gpuweb#3435, by @jiangzhaoming.
-
Object labels require client-shared state gpuweb/gpuweb#3263, by @kainino0x.
-
Conversion expressions for the Abstract types gpuweb/gpuweb#2959, by @dneto0.
-
WGSL pointer aliasing rules gpuweb/gpuweb#1457, by @dneto0.
-
-
57 issues received 157 new comments.
-
wgsl: pointer-to-workgroup in user-defined functions gpuweb/gpuweb#3276 (15 comments by @alan-baker, @dneto0, @kdashg, @litherum)
-
Canvas texture expiry timing is unpredictable gpuweb/gpuweb#3295 (11 comments by @greggman, @kainino0x, @litherum, @toji)
-
createPipelineAsync() doesn't reject its promise, even if the created pipeline is invalid gpuweb/gpuweb#3296 (10 comments by @greggman, @kainino0x, @Kangz, @kdashg, @litherum)
-
Ray Tracing extension gpuweb/gpuweb#535 (10 comments by @expenses, @Kangz, @litherum, @magcius, @maierfelix, @MatusT)
-
Override constant should not provided by pipeline as NaN or Inf gpuweb/gpuweb#3435 (8 comments by @jiangzhaoming, @kainino0x, @Kangz, @kdashg)
-
Does getBindGroupLayout return the same underlying object or a new one? gpuweb/gpuweb#3454 (6 comments by @greggman, @kainino0x, @Kangz)
-
wgsl: atan2 result depends on signs of both arguments, not spec'd gpuweb/gpuweb#3437 (5 comments by @dneto0, @zoddicus)
-
Uniformity analysis is far too strict gpuweb/gpuweb#3479 (4 comments by @brendan-duncan, @jimblandy, @Kangz)
-
Object labels require client-shared state gpuweb/gpuweb#3263 (4 comments by @kainino0x, @Kangz, @litherum)
-
`textureSampleBaseClampToEdge` clamping should clamp a half-texel from edge gpuweb/gpuweb#3471 (4 comments by @dneto0, @kainino0x)
-
wgsl: add way to get special values for a given numeric type (e.g. max, min, lowest) gpuweb/gpuweb#3431 (4 comments by @dneto0, @kdashg)
-
Inconsistent status of leading zeros gpuweb/gpuweb#3415 (4 comments by @alan-baker, @kdashg, @mehmetoguzderin)
-
Clarify intended behaviour when packing out of range values gpuweb/gpuweb#3470 (3 comments by @dneto0, @zoddicus)
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766 (3 comments by @dneto0, @kainino0x)
-
Clarification needed with regards to flushing denorm f16 values gpuweb/gpuweb#3421 (3 comments by @alan-baker, @kdashg, @litherum)
-
Add precedence rules description into spec gpuweb/gpuweb#3346 (3 comments by @dj2, @jiangzhaoming, @kdashg)
-
mapSync on Workers - and possibly on the main thread gpuweb/gpuweb#2217 (2 comments by @lisanlow, @toji)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (2 comments by @r12a, @xfq)
-
GPURenderBundleEncoder.setPipeline() not implementable on MTLIndirectCommandBuffer gpuweb/gpuweb#3423 (2 comments by @Kangz, @litherum)
-
Colorspace conversions for GPUExternalTexture are arbitrarily complex gpuweb/gpuweb#3293 (2 comments by @kainino0x, @kdashg)
-
Specification needs to clarify behavior of indexing of abstract composite type with runtime value gpuweb/gpuweb#3210 (2 comments by @dneto0, @kdashg)
-
why can't *any* fixed-size array use an element count that is an override expression gpuweb/gpuweb#3475 (1 comment by @dneto0)
-
Clarify storable type concrete components for vectors and matrices. gpuweb/gpuweb#3031 (1 comment by @dneto0)
-
Add text and example around IO locations spanning multiple locations gpuweb/gpuweb#3045 (1 comment by @dneto0)
-
Update `Pipeline input and output interface` section to reference `@builtin` and `@location` gpuweb/gpuweb#2996 (1 comment by @dneto0)
-
Add texture types to `t` for textureSampleGrad and textureSampleBias gpuweb/gpuweb#2892 (1 comment by @dneto0)
-
sampleType: "float" is not valid with multisampled: true gpuweb/gpuweb#3364 (1 comment by @kainino0x)
-
Difference between `src-alpha` and `src-alpha-saturated` gpuweb/gpuweb#3432 (1 comment by @kainino0x)
-
How do we specify the timing of external texture expiration? gpuweb/gpuweb#3324 (1 comment by @kdashg)
-
wgsl: Add limits to avoid nuisance floating point literals gpuweb/gpuweb#3255 (1 comment by @kdashg)
-
wgsl: compute shaders have no user-defined inputs or outputs gpuweb/gpuweb#3227 (1 comment by @alan-baker)
-
WGSL spec gives readers the runaround regarding `var` access modes gpuweb/gpuweb#3419 (1 comment by @alan-baker)
-
invalid expressions in const-expr and override-expr should be error (no later than pipeline creation time) gpuweb/gpuweb#3253 (1 comment by @kdashg)
-
Numeric coercion leads to surprising results with division gpuweb/gpuweb#3326 (1 comment by @kdashg)
-
Clarify meaning of `context` in integer example in `TYPE INFERENCE FOR LITERALS` gpuweb/gpuweb#2970 (1 comment by @kdashg)
-
Definition of length(x) is too restrictive for x being a scalar gpuweb/gpuweb#3407 (1 comment by @dneto0)
-
-
49 pull requests opened.
-
Clarify definition of the concretization of a type. gpuweb/gpuweb#3482, by @jimblandy.
-
Validation errors on out-of-bounds pipeline constants, and clearValues gpuweb/gpuweb#3481, by @kainino0x.
-
Streamline tokens gpuweb/gpuweb#3480, by @mehmetoguzderin.
-
Require effective storage buffer binding size be a multiple of 4 gpuweb/gpuweb#3477, by @Jiawei-Shao.
-
Rewwrite atan2 definition gpuweb/gpuweb#3474, by @dneto0.
-
textureSampelBaseClampToEdge clamps to half-texel from the edge gpuweb/gpuweb#3472, by @dneto0.
-
Allow u32s for attributes gpuweb/gpuweb#3469, by @alan-baker.
-
Explain why u32(AbstractInt) conversion expression exists gpuweb/gpuweb#3467, by @dneto0.
-
Further qualify composite types as decomposible gpuweb/gpuweb#3466, by @dneto0.
-
Explain duality of ldexp with frexp gpuweb/gpuweb#3464, by @dneto0.
-
Restore arithmetic right shift, when LHS is signed gpuweb/gpuweb#3463, by @dneto0.
-
[editorial] Type chapter clarifications gpuweb/gpuweb#3462, by @alan-baker.
-
Make t descriptions consistent for texureSampleBias textureSampleGrad (take2) gpuweb/gpuweb#3460, by @dneto0.
-
Sign change error should be for << gpuweb/gpuweb#3459, by @alan-baker.
-
Add missing override-declaration to uniformity rules for lvalue-position gpuweb/gpuweb#3456, by @dneto0.
-
Revert "Make getBGL return a "copy" instead of a "reference"" gpuweb/gpuweb#3455, by @kainino0x.
-
Make t descriptions consistent for texureSampleBias textureSampleGrad gpuweb/gpuweb#3452, by @dneto0.
-
Fix t parameterization of textureSampleLevel gpuweb/gpuweb#3451, by @dneto0.
-
Clarify for-loop condition evaluation gpuweb/gpuweb#3450, by @dneto0.
-
Fix t parameterization of textureSample gpuweb/gpuweb#3449, by @dneto0.
-
Clarify ranges of normalized values for unorm, snorm packing/unpacking gpuweb/gpuweb#3448, by @dneto0.
-
Clarify flushing of conversions between different width float types gpuweb/gpuweb#3447, by @dneto0.
-
Clarify abstract shift left gpuweb/gpuweb#3446, by @dneto0.
-
Fix ldexp constraints lost when SPIR-V mappings were removed gpuweb/gpuweb#3445, by @dneto0.
-
wgsl: Rename "array_type" to "array_type_specifier" gpuweb/gpuweb#3444, by @dneto0.
-
Fix WGSL spec build error gpuweb/gpuweb#3443, by @mehmetoguzderin.
-
wgsl: clarify type equivalence of arrays with overridable element counts gpuweb/gpuweb#3442, by @dneto0.
-
wgsl: Fix loop example (let should be var) gpuweb/gpuweb#3440, by @dneto0.
-
wgsl: Fix link to WebGPU GPUExternalTexture gpuweb/gpuweb#3439, by @dneto0.
-
Clarify initial description if, else-if, else gpuweb/gpuweb#3438, by @dneto0.
-
Rename `type_decl` to `type_specifier` gpuweb/gpuweb#3436, by @mehmetoguzderin.
-
Rename maxColorAttachmentBytesPerPixel to *PerSample gpuweb/gpuweb#3434, by @kainino0x.
-
nit: only resolves are ordered, not rejections gpuweb/gpuweb#3433, by @kainino0x.
-
Add HTMLImageElement and ImageData to copyExternalImageToTexture gpuweb/gpuweb#3430, by @kainino0x.
-
[editorial] Change pipeline IO to shader-stage IO gpuweb/gpuweb#3429, by @alan-baker.
-
Change atan2's accuracy to being inherited from atan(y/x) gpuweb/gpuweb#3428, by @zoddicus.
-
Include WGSL chair in contacts gpuweb/gpuweb#3427, by @mehmetoguzderin.
-
[editorial] Fix access mode defaults gpuweb/gpuweb#3426, by @alan-baker.
-
Constant floating point errors for NaN and inf gpuweb/gpuweb#3425, by @alan-baker.
-
wgsl: ban ptr-to-workgroup params to user-defined functions gpuweb/gpuweb#3422, by @dneto0.
-
Remove old system config gpuweb/gpuweb#3417, by @mehmetoguzderin.
-
Synchronize grammar dependencies for new version gpuweb/gpuweb#3416, by @mehmetoguzderin.
-
Adjust literals section gpuweb/gpuweb#3414, by @mehmetoguzderin.
-
Upgrade dependency for grammar and ignore pre gpuweb/gpuweb#3413, by @mehmetoguzderin.
-
Add HTMLVideoElement to copyExternalImageToTexture gpuweb/gpuweb#3410, by @kainino0x.
-
wgsl: clarify length builtin on scalar gpuweb/gpuweb#3409, by @dneto0.
-
Matching rules for fragment shader output and color attachments gpuweb/gpuweb#3408, by @kainino0x.
-
Use CDN for Preview URLs instead of Pages gpuweb/gpuweb#3406, by @mehmetoguzderin.
-
[editorial] "create a WebGPU object" algorithm gpuweb/gpuweb#3405, by @kainino0x.
-
-
46 pull requests merged.
-
Validation errors on out-of-bounds pipeline constants, and clearValues gpuweb/gpuweb#3481.
-
Allow u32s for attributes gpuweb/gpuweb#3469.
-
textureSampelBaseClampToEdge clamps to half-texel from the edge gpuweb/gpuweb#3472.
-
Further qualify composite types as decomposible gpuweb/gpuweb#3466.
-
Explain why u32(AbstractInt) conversion expression exists gpuweb/gpuweb#3467.
-
Explain duality of ldexp with frexp gpuweb/gpuweb#3464.
-
Fix t parameterization of textureSampleLevel gpuweb/gpuweb#3451.
-
[editorial] Type chapter clarifications gpuweb/gpuweb#3462.
-
Restore arithmetic right shift, when LHS is signed gpuweb/gpuweb#3463.
-
Add a static alias analysis gpuweb/gpuweb#3336.
-
Sign change error should be for << gpuweb/gpuweb#3459.
-
Clarify flushing of conversions between different width float types gpuweb/gpuweb#3447.
-
Clarify ranges of normalized values for unorm, snorm packing/unpacking gpuweb/gpuweb#3448.
-
Clarify for-loop condition evaluation gpuweb/gpuweb#3450.
-
Add missing override-declaration to uniformity rules for lvalue-position gpuweb/gpuweb#3456.
-
Revert "Make getBGL return a "copy" instead of a "reference"" gpuweb/gpuweb#3455.
-
Clarify abstract shift left gpuweb/gpuweb#3446.
-
Fix t parameterization of textureSample gpuweb/gpuweb#3449.
-
wgsl: Rename "array_type" to "array_type_specifier" gpuweb/gpuweb#3444.
-
Fix ldexp constraints lost when SPIR-V mappings were removed gpuweb/gpuweb#3445.
-
Fix WGSL spec build error gpuweb/gpuweb#3443.
-
Rename `type_decl` to `type_specifier` gpuweb/gpuweb#3436.
-
wgsl: clarify type equivalence of arrays with overridable element counts gpuweb/gpuweb#3442.
-
Describe who can contribute and the CG/WG relation gpuweb/gpuweb#3342.
-
Describe validation errors from createShaderModule gpuweb/gpuweb#3347.
-
wgsl: Fix loop example (let should be var) gpuweb/gpuweb#3440.
-
wgsl: Fix link to WebGPU GPUExternalTexture gpuweb/gpuweb#3439.
-
Clarify initial description if, else-if, else gpuweb/gpuweb#3438.
-
Add texture sample function which clamps coordinates gpuweb/gpuweb#3403.
-
nit: only resolves are ordered, not rejections gpuweb/gpuweb#3433.
-
Adjust literals section gpuweb/gpuweb#3414.
-
Constant floating point errors for NaN and inf gpuweb/gpuweb#3425.
-
[editorial] Change pipeline IO to shader-stage IO gpuweb/gpuweb#3429.
-
Change atan2's accuracy to being inherited from atan(y/x) gpuweb/gpuweb#3428.
-
Include WGSL chair in contacts gpuweb/gpuweb#3427.
-
[editorial] Uniformity examples gpuweb/gpuweb#3379.
-
Add HTMLVideoElement to copyExternalImageToTexture gpuweb/gpuweb#3410.
-
[editorial] Fix access mode defaults gpuweb/gpuweb#3426.
-
Constant errors for integer expressions gpuweb/gpuweb#3401.
-
wgsl: clarify length builtin on scalar gpuweb/gpuweb#3409.
-
Remove old system config gpuweb/gpuweb#3417.
-
Synchronize grammar dependencies for new version gpuweb/gpuweb#3416.
-
Upgrade dependency for grammar and ignore pre gpuweb/gpuweb#3413.
-
[editorial] "create a WebGPU object" algorithm gpuweb/gpuweb#3405.
-
No-op unmap on lost device, related nits gpuweb/gpuweb#3404.
-
Use CDN for Preview URLs instead of Pages gpuweb/gpuweb#3406.
-
-
4 pull requests closed.
-
Make t descriptions consistent for texureSampleBias textureSampleGrad gpuweb/gpuweb#3452.
-
Make t descriptions consistent for texureSampleBias textureSampleGrad (take2) gpuweb/gpuweb#3460.
-
Make rg11b10ufloat RENDER_ATTACHMENT:yes and resolve:yes. gpuweb/gpuweb#2659.
-
Make GPUExternalTexture sampling always use clamping address mode. gpuweb/gpuweb#2796.
-
-
43 pull requests received 161 new comments.
-
Rename `type_decl` to `type_specifier` gpuweb/gpuweb#3436 (41 comments by @dj2, @dneto0, @mehmetoguzderin)
-
Clarify ranges of normalized values for unorm, snorm packing/unpacking gpuweb/gpuweb#3448 (8 comments by @alan-baker, @dj2, @dneto0)
-
mapAsync(WRITE) and zero-filling gpuweb/gpuweb#2926 (7 comments by @kainino0x, @Kangz, @kdashg, @litherum)
-
Make t descriptions consistent for texureSampleBias textureSampleGrad gpuweb/gpuweb#3452 (6 comments by @alan-baker, @dneto0)
-
Clarify flushing of conversions between different width float types gpuweb/gpuweb#3447 (6 comments by @alan-baker, @dneto0)
-
Upgrade dependency for grammar and ignore pre gpuweb/gpuweb#3413 (6 comments by @kainino0x, @mehmetoguzderin)
-
Require effective storage buffer binding size be a multiple of 4 gpuweb/gpuweb#3477 (4 comments by @kainino0x, @Kangz)
-
Restore arithmetic right shift, when LHS is signed gpuweb/gpuweb#3463 (4 comments by @amaiorano, @dneto0)
-
wgsl: clarify type equivalence of arrays with overridable element counts gpuweb/gpuweb#3442 (4 comments by @alan-baker, @dneto0)
-
Add texture sample function which clamps coordinates gpuweb/gpuweb#3403 (4 comments by @dneto0, @kainino0x, @kdashg)
-
Describe validation errors from createShaderModule gpuweb/gpuweb#3347 (3 comments by @dneto0, @kainino0x)
-
[editorial] Change pipeline IO to shader-stage IO gpuweb/gpuweb#3429 (3 comments by @dneto0, @mehmetoguzderin)
-
Validation errors on out-of-bounds pipeline constants, and clearValues gpuweb/gpuweb#3481 (2 comments by @kainino0x, @toji)
-
Describe who can contribute and the CG/WG relation gpuweb/gpuweb#3342 (2 comments by @kainino0x, @tidoust)
-
Add HTMLImageElement and ImageData to copyExternalImageToTexture gpuweb/gpuweb#3430 (2 comments by @kainino0x)
-
Change atan2's accuracy to being inherited from atan(y/x) gpuweb/gpuweb#3428 (2 comments by @alan-baker, @zoddicus)
-
Make t descriptions consistent for texureSampleBias textureSampleGrad (take2) gpuweb/gpuweb#3460 (1 comment by @dneto0)
-
Revert "Make getBGL return a "copy" instead of a "reference"" gpuweb/gpuweb#3455 (1 comment by @kainino0x)
-
Make rg11b10ufloat RENDER_ATTACHMENT:yes and resolve:yes. gpuweb/gpuweb#2659 (1 comment by @kainino0x)
-
wgsl: ban ptr-to-workgroup params to user-defined functions gpuweb/gpuweb#3422 (1 comment by @kdashg)
-
Matching rules for fragment shader output and color attachments gpuweb/gpuweb#3408 (1 comment by @Kangz)
-
Make GPUExternalTexture sampling always use clamping address mode. gpuweb/gpuweb#2796 (1 comment by @kainino0x)
-
TR publications
-
30 documents published.
-
WebGPU published 15 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 15 times as a Working Draft. History of WebGPU Shading Language
-
Participants
-
1 participant joined.
-
Yuan Yao (ByteDance) joined.
-
-
1 participant left.
-
Dzmitry Malyshau (Mozilla Foundation) left.
-
GitHub
gpuweb/gpuweb
-
39 issues opened.
-
[editorial] unmap should check for device loss gpuweb/gpuweb#3402, by @kainino0x.
-
wgsl: Clarify whether non-entry-point functions can take arguments with locations gpuweb/gpuweb#3399, by @nical.
-
Confusing text around override array size gpuweb/gpuweb#3390, by @dj2.
-
Definition of `hex_float_literal` seems incorrect gpuweb/gpuweb#3389, by @vincentisambart.
-
preview-pull-request doesn't pick up new diagrams gpuweb/gpuweb#3386, by @kainino0x.
-
Clarify valid address spaces for runtime-sized arrays gpuweb/gpuweb#3383, by @apbodnar.
-
Runtime behavior of NaN-producing math functions gpuweb/gpuweb#3365, by @litherum.
-
sampleType: "float" is not valid with multisampled: true gpuweb/gpuweb#3364, by @kainino0x.
-
Clarify references to depth24unorm gpuweb/gpuweb#3359, by @kainino0x.
-
add AbstractFloat overloads for modf and frexp? gpuweb/gpuweb#3358, by @dneto0.
-
Why does CopyExternalImageToTexture() require RENDER_ATTACHMENT usage? gpuweb/gpuweb#3357, by @shi-yan.
-
wgsl: rename type_decl grammar element to type_specifier gpuweb/gpuweb#3356, by @dneto0.
-
item size support in array gpuweb/gpuweb#3355, by @mikialex.
-
GPURenderPassDescriptor.maxDrawCount seems arcane gpuweb/gpuweb#3349, by @juj.
-
Add precedence rules description into spec gpuweb/gpuweb#3346, by @dj2.
-
Adding support for rgba16unorm texture format gpuweb/gpuweb#3333, by @ankit040683.
-
editorial: set label on all created objects gpuweb/gpuweb#3332, by @kainino0x.
-
Use cube texture as a storage type gpuweb/gpuweb#3331, by @JonahPlusPlus.
-
`stripIndexFormat` might be a surprising pipeline-creation-time requirement. gpuweb/gpuweb#3328, by @kdashg.
-
Numeric coercion leads to surprising results with division gpuweb/gpuweb#3326, by @jimblandy.
-
How do we specify the timing of external texture expiration? gpuweb/gpuweb#3324, by @kainino0x.
-
wgsl: atanh and friends are "inverse hyperbolic ***" or "area hyperbolic **" not "arc ***" gpuweb/gpuweb#3323, by @dneto0.
-
How to enable texture transparency ? gpuweb/gpuweb#3317, by @jay19240.
-
create*PipelineAsync prevents using the pipeline synchronously gpuweb/gpuweb#3297, by @kainino0x.
-
createPipelineAsync() doesn't reject its promise, even if the created pipeline is invalid gpuweb/gpuweb#3296, by @litherum.
-
Canvas texture expiry timing is unpredictable gpuweb/gpuweb#3295, by @kainino0x.
-
wgsl: limit nesting depth of switch statements gpuweb/gpuweb#3294, by @dneto0.
-
Colorspace conversions for GPUExternalTexture are arbitrarily complex gpuweb/gpuweb#3293, by @litherum.
-
Accuracy of `acosh` is does not include Vulkan CTS alternative gpuweb/gpuweb#3287, by @zoddicus.
-
wgsl: publish a form of the grammar suitable for recursive descent parsing gpuweb/gpuweb#3286, by @dneto0.
-
`cosh` and `sinh` have the same inherited accuracy gpuweb/gpuweb#3281, by @zoddicus.
-
wgsl: describe how the name `vec4` in `vec4<f32>` is resolved gpuweb/gpuweb#3280, by @dneto0.
-
Lower the max binding index in createBindGroupLayout gpuweb/gpuweb#3279, by @nical.
-
Remove the `Type-defining Keywords` section gpuweb/gpuweb#3277, by @dj2.
-
wgsl: pointer-to-workgroup in user-defined functions gpuweb/gpuweb#3276, by @dneto0.
-
Editorial: Out-of-bounds accesses can occur without interacting with memory gpuweb/gpuweb#3273, by @litherum.
-
Accessing an "invalid memory reference": Trapping behavior gpuweb/gpuweb#3272, by @litherum.
-
`GPUBuffer`s shouldn't have to accumulate a list of `mapAsync` promises to reject gpuweb/gpuweb#3271, by @jimblandy.
-
-
66 issues closed.
-
remove keywords that are context-dependent names gpuweb/gpuweb#3166.
-
Expose WGSL errors in the API gpuweb/gpuweb#2308.
-
`stripIndexFormat` might be a surprising pipeline-creation-time requirement. gpuweb/gpuweb#3328.
-
Consistent definitions of the methods gpuweb/gpuweb#760.
-
Consider making ShaderModule compilation errors special gpuweb/gpuweb#2119.
-
Specify the units for the various `cos`, `sin`, etc builtin functions gpuweb/gpuweb#3191.
-
Numeric coercion leads to surprising results with division gpuweb/gpuweb#3326.
-
Definition of `hex_float_literal` seems incorrect gpuweb/gpuweb#3389.
-
wgsl: Other language types in reserved word list gpuweb/gpuweb#2983.
-
Buffer state validation in unmap() is unnecessary gpuweb/gpuweb#3251.
-
Clarify valid address spaces for runtime-sized arrays gpuweb/gpuweb#3383.
-
preview-pull-request doesn't pick up new diagrams gpuweb/gpuweb#3386.
-
Lower the max binding index in createBindGroupLayout gpuweb/gpuweb#3279.
-
GPURenderPassDescriptor.maxDrawCount seems arcane gpuweb/gpuweb#3349.
-
Why does CopyExternalImageToTexture() require RENDER_ATTACHMENT usage? gpuweb/gpuweb#3357.
-
maxColorAttachments not sufficient? gpuweb/gpuweb#2965.
-
Limit for the maximum buffer size gpuweb/gpuweb#1371.
-
Out of memory handling for GPUBuffer mapping gpuweb/gpuweb#872.
-
wgsl: publish a form of the grammar suitable for recursive descent parsing gpuweb/gpuweb#3286.
-
Ordering of promise resolution. gpuweb/gpuweb#962.
-
Clarify references to depth24unorm gpuweb/gpuweb#3359.
-
Multisample resolve when base texture formats do not match gpuweb/gpuweb#2729.
-
wgsl: permit certain resource variables to overlap gpuweb/gpuweb#3266.
-
GPUBuffer.mapAsync is overly pessimistic gpuweb/gpuweb#2646.
-
Annotate parts of the spec as "Tracking Vectors" gpuweb/gpuweb#950.
-
Improved style for issuing work between timelines gpuweb/gpuweb#2961.
-
Try-catching getMappedRange is useful but shouldn't be gpuweb/gpuweb#3013.
-
Is an entry point required in WGSL? gpuweb/gpuweb#3130.
-
Confusing languages in technical overview around conversions. gpuweb/gpuweb#2910.
-
Use cube texture as a storage type gpuweb/gpuweb#3331.
-
[editorial] Make object creation to respect the device timeline gpuweb/gpuweb#1970.
-
wgsl: limit layout size of a type gpuweb/gpuweb#2118.
-
Uniformity analysis of local variables gpuweb/gpuweb#2859.
-
Make GPUDevice.destroy() a queued operation gpuweb/gpuweb#2599.
-
How to enable texture transparency ? gpuweb/gpuweb#3317.
-
wgsl: atomic type section should link to memory model section gpuweb/gpuweb#2978.
-
atomic types section references `QueueFamily` which doesn't seem to be defined gpuweb/gpuweb#2979.
-
Replace instances of `integral` with `integer scalar` gpuweb/gpuweb#3133.
-
Clarification on control flow of function calls gpuweb/gpuweb#3198.
-
Clarify function parameter initialization gpuweb/gpuweb#3196.
-
Clarification on parameter matching statement gpuweb/gpuweb#3197.
-
Update links for user-defined functions gpuweb/gpuweb#3194.
-
wgsl: limit nesting depth of switch statements gpuweb/gpuweb#3294.
-
Clarification around `vector` and `matrix` for creation-time constant gpuweb/gpuweb#3074.
-
Make definition of `invalid memory reference` more generic gpuweb/gpuweb#3128.
-
Allow optional `()`s on `enable`? gpuweb/gpuweb#3207.
-
Clarification in user-defined functions on attributes on parameters gpuweb/gpuweb#3193.
-
`workgroup_size` valid values gpuweb/gpuweb#2915.
-
wsgl: disallow repeated use of builtin inputs and builtin outputs gpuweb/gpuweb#3236.
-
timestamp-query is unimplementable on TBDR architectures gpuweb/gpuweb#2046.
-
Drop support for macOS 10.12? gpuweb/gpuweb#3238.
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094.
-
`cosh` and `sinh` have the same inherited accuracy gpuweb/gpuweb#3281.
-
Clarification of `@builtin` values in structs not used in an entry-point gpuweb/gpuweb#3261.
-
Clarification of loop identifier lifetime gpuweb/gpuweb#3153.
-
Entry point declaration example bug gpuweb/gpuweb#3202.
-
User defined function example typo gpuweb/gpuweb#3195.
-
Duplicate text in enable section gpuweb/gpuweb#3216.
-
Buffer Offset issue in buffer-texture copies with depth stencil textures gpuweb/gpuweb#3185.
-
1 issue reopened.
-
GPUBuffer.mapAsync is overly pessimistic gpuweb/gpuweb#2646, by @kainino0x.
-
-
75 issues received 341 new comments.
-
wgsl: publish a form of the grammar suitable for recursive descent parsing gpuweb/gpuweb#3286 (36 comments by @dj2, @dneto0)
-
WGSL pointer aliasing rules gpuweb/gpuweb#1457 (27 comments by @alan-baker, @dneto0, @jimblandy, @kdashg, @krogovin, @litherum)
-
invalid expressions in const-expr and override-expr should be error (no later than pipeline creation time) gpuweb/gpuweb#3253 (19 comments by @dneto0, @jimblandy, @kdashg, @litherum)
-
Canvas texture expiry timing is unpredictable gpuweb/gpuweb#3295 (19 comments by @kainino0x, @kenrussell, @litherum, @toji)
-
Accessing an "invalid memory reference": Trapping behavior gpuweb/gpuweb#3272 (17 comments by @alan-baker, @dneto0, @kdashg, @litherum)
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766 (14 comments by @kainino0x, @kdashg, @litherum)
-
wgsl: limit nesting depth of switch statements gpuweb/gpuweb#3294 (13 comments by @dj2, @dneto0, @kdashg, @litherum)
-
Colorspace conversions for GPUExternalTexture are arbitrarily complex gpuweb/gpuweb#3293 (12 comments by @bjjones, @kainino0x, @Kangz, @kdashg, @litherum, @shaoboyan)
-
`GPUBuffer`s shouldn't have to accumulate a list of `mapAsync` promises to reject gpuweb/gpuweb#3271 (12 comments by @jimblandy, @kainino0x, @Kangz, @nical)
-
preview-pull-request doesn't pick up new diagrams gpuweb/gpuweb#3386 (11 comments by @dj2, @kainino0x, @mehmetoguzderin)
-
importExternalTexture/copyExternalImageToTexture should not throw when source image is not decodable gpuweb/gpuweb#3288 (8 comments by @kainino0x, @litherum)
-
Accuracy of `acosh` is does not include Vulkan CTS alternative gpuweb/gpuweb#3287 (8 comments by @dneto0, @zoddicus)
-
Precision requirements for `atan2` appear to be too restrictive for some backends gpuweb/gpuweb#3209 (7 comments by @jimblandy, @kdashg, @litherum, @munrocket)
-
Buffer state validation in unmap() is unnecessary gpuweb/gpuweb#3251 (6 comments by @kainino0x, @kdashg)
-
create*PipelineAsync prevents using the pipeline synchronously gpuweb/gpuweb#3297 (6 comments by @kainino0x, @toji)
-
`stripIndexFormat` might be a surprising pipeline-creation-time requirement. gpuweb/gpuweb#3328 (5 comments by @kainino0x, @kdashg)
-
Numeric coercion leads to surprising results with division gpuweb/gpuweb#3326 (4 comments by @ben-clayton, @dneto0, @jimblandy)
-
wgsl: permit certain resource variables to overlap gpuweb/gpuweb#3266 (4 comments by @kdashg, @litherum)
-
Change ArrayBuffer arguments to ArrayBuffer + srcOffset gpuweb/gpuweb#244 (4 comments by @juj, @kainino0x)
-
wgsl: language evolution: managing rollout of core (essentially sugar) language features gpuweb/gpuweb#3149 (4 comments by @kdashg)
-
Lower the max binding index in createBindGroupLayout gpuweb/gpuweb#3279 (4 comments by @kainino0x, @kdashg, @nical)
-
timestamp-query is unimplementable on TBDR architectures gpuweb/gpuweb#2046 (4 comments by @haoxli, @Kangz, @litherum)
-
wgsl: Clarify whether non-entry-point functions can take arguments with locations gpuweb/gpuweb#3399 (3 comments by @dj2, @nical)
-
Definition of `hex_float_literal` seems incorrect gpuweb/gpuweb#3389 (3 comments by @dneto0, @mehmetoguzderin, @vincentisambart)
-
How do we specify the timing of external texture expiration? gpuweb/gpuweb#3324 (3 comments by @kainino0x)
-
Object labels require client-shared state gpuweb/gpuweb#3263 (3 comments by @kainino0x, @kdashg, @toji)
-
wgsl: describe how the name `vec4` in `vec4<f32>` is resolved gpuweb/gpuweb#3280 (3 comments by @dneto0, @munrocket)
-
wgsl: Add limits to avoid nuisance floating point literals gpuweb/gpuweb#3255 (3 comments by @dneto0, @kdashg)
-
Distinction between content-timeline and device-timeline state gpuweb/gpuweb#2962 (2 comments by @kainino0x)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (2 comments by @kainino0x, @Kangz)
-
Runtime behavior of NaN-producing math functions gpuweb/gpuweb#3365 (2 comments by @alan-baker, @dneto0)
-
Multisample resolve when base texture formats do not match gpuweb/gpuweb#2729 (2 comments by @austinEng, @kainino0x)
-
Why does CopyExternalImageToTexture() require RENDER_ATTACHMENT usage? gpuweb/gpuweb#3357 (2 comments by @shaoboyan, @shi-yan)
-
wgsl: Missing specification on modf behavior for negatives. gpuweb/gpuweb#2468 (2 comments by @brent-williams, @stegu)
-
wgsl: Add some clarification in the `Scalar Types` section that an Abstract is not a scalar type gpuweb/gpuweb#2975 (2 comments by @alan-baker, @dj2)
-
createPipelineAsync() doesn't reject its promise, even if the created pipeline is invalid gpuweb/gpuweb#3296 (2 comments by @kainino0x)
-
Add text and example around IO locations spanning multiple locations gpuweb/gpuweb#3045 (2 comments by @dj2)
-
"validating linear texture data" should depend on the aspect for depth-stencil formats gpuweb/gpuweb#2569 (1 comment by @kainino0x)
-
Add a note that copyT2T might not preserve the exact bit values gpuweb/gpuweb#2558 (1 comment by @kainino0x)
-
Make lost devices appear to function as much as possible gpuweb/gpuweb#1629 (1 comment by @kainino0x)
-
Possibility of SPIR-V and/or GLSL as a WebGPU extension? gpuweb/gpuweb#847 (1 comment by @kevincrans)
-
Accuracy of `mix` does not cover all common implementations of linear blend gpuweb/gpuweb#3260 (1 comment by @kdashg)
-
List of issues to discuss with TC39, WASM, JS engine teams gpuweb/gpuweb#747 (1 comment by @kainino0x)
-
[editorial] Make object creation to respect the device timeline gpuweb/gpuweb#1970 (1 comment by @kainino0x)
-
Clarification around `vector` and `matrix` for creation-time constant gpuweb/gpuweb#3074 (1 comment by @alan-baker)
-
wsgl: disallow repeated use of builtin inputs and builtin outputs gpuweb/gpuweb#3236 (1 comment by @kdashg)
-
Buffer Offset issue in buffer-texture copies with depth stencil textures gpuweb/gpuweb#3185 (1 comment by @Kangz)
-
Editorial: Out-of-bounds accesses can occur without interacting with memory gpuweb/gpuweb#3273 (1 comment by @alan-baker)
-
-
87 pull requests opened.
-
No-op unmap on lost device, related nits gpuweb/gpuweb#3404, by @kainino0x.
-
Add texture sample function which clamps coordinates gpuweb/gpuweb#3403, by @kainino0x.
-
Constant errors for integer expressions gpuweb/gpuweb#3401, by @alan-baker.
-
Errors for constant out-of-bounds accesses gpuweb/gpuweb#3400, by @alan-baker.
-
[editorial] consistent x-refs to IDL methods, add a common dfn for setBindGroup gpuweb/gpuweb#3398, by @kainino0x.
-
Add note about texture copies not preserving bit values gpuweb/gpuweb#3397, by @kainino0x.
-
[editorial] Clarify specification of stripIndexFormat gpuweb/gpuweb#3396, by @kainino0x.
-
Remove 'invariant' from reserved word list gpuweb/gpuweb#3395, by @dneto0.
-
Pull reserved word list into its own file gpuweb/gpuweb#3394, by @dneto0.
-
Preview includes gpuweb/gpuweb#3393, by @mehmetoguzderin.
-
Clarify units for trig and hyperbolic functions gpuweb/gpuweb#3392, by @dneto0.
-
[editorial] highlight data-link-type=idl too gpuweb/gpuweb#3391, by @kainino0x.
-
Fix address space for runtime-sized array example gpuweb/gpuweb#3388, by @dneto0.
-
Fix PR previews with diagrams gpuweb/gpuweb#3387, by @kainino0x.
-
Fix declaration of [SameObject]-ness on interface rather than dictionary gpuweb/gpuweb#3385, by @dontcallmedom.
-
Deploy previews to GitHub gpuweb/gpuweb#3384, by @mehmetoguzderin.
-
[editorial] Define scheme for state on different timelines gpuweb/gpuweb#3382, by @kainino0x.
-
[editorial] Make `GPUUncapturedErrorEvent.error` `[SameObject]` gpuweb/gpuweb#3381, by @kainino0x.
-
Require maxBufferSize to be at least as max*BufferBindingSize gpuweb/gpuweb#3380, by @kainino0x.
-
[editorial] Uniformity examples gpuweb/gpuweb#3379, by @alan-baker.
-
Move promise resolutions in algorithms to content timeline gpuweb/gpuweb#3378, by @toji.
-
[editorial] Fix logo's gpuweb/gpuweb#3377, by @munrocket.
-
Style links with the timeline of the reference they point to gpuweb/gpuweb#3375, by @toji.
-
[editorial] Improve exposition of the uniformity analysis gpuweb/gpuweb#3374, by @alan-baker.
-
Rename 'Creation' headers for disambiguation gpuweb/gpuweb#3373, by @toji.
-
[editorial] Fix uniformity for statements table render gpuweb/gpuweb#3372, by @alan-baker.
-
Add timestampWrites to the section of Timestamp Query gpuweb/gpuweb#3371, by @haoxli.
-
Move timestamp query inside passes into a separate extension doc gpuweb/gpuweb#3370, by @haoxli.
-
[editorial] Vertical column headers for texture format table gpuweb/gpuweb#3369, by @kainino0x.
-
Remove device-timeline validation error in unmap() gpuweb/gpuweb#3368, by @kainino0x.
-
Add per-pixel accounting of render target depths gpuweb/gpuweb#3367, by @litherum.
-
Add maxBufferSize limit gpuweb/gpuweb#3366, by @litherum.
-
Require render and resolve to have same base format gpuweb/gpuweb#3363, by @kainino0x.
-
Make error reporting unordered gpuweb/gpuweb#3362, by @kainino0x.
-
Define promise resolution ordering gpuweb/gpuweb#3361, by @kainino0x.
-
[editorial] Clarify 24-bit depth gpuweb/gpuweb#3360, by @kainino0x.
-
[editorial] Specify what part of buffer is updated in unmap() gpuweb/gpuweb#3353, by @kainino0x.
-
[editorial] Fix logo's gpuweb/gpuweb#3352, by @munrocket.
-
Specify OOM behavior in mapping gpuweb/gpuweb#3351, by @kainino0x.
-
Update the WebGPU editors list gpuweb/gpuweb#3350, by @kainino0x.
-
Reject redundant map requests early so there are never multiple pending promises gpuweb/gpuweb#3348, by @kainino0x.
-
Describe validation errors from createShaderModule gpuweb/gpuweb#3347, by @kainino0x.
-
[editorial] Fix attribute typos gpuweb/gpuweb#3345, by @alan-baker.
-
[editorial] Introduction clarifications gpuweb/gpuweb#3344, by @alan-baker.
-
[Editorial] Clarifying Abstract Types in the first line gpuweb/gpuweb#3343, by @munrocket.
-
Describe who can contribute and the CG/WG relation gpuweb/gpuweb#3342, by @kainino0x.
-
Make getBGL return a "copy" instead of a "reference" gpuweb/gpuweb#3341, by @kainino0x.
-
Add tracking-vector annotations gpuweb/gpuweb#3340, by @kainino0x.
-
Better formalize and document mapAsync promise ordering gpuweb/gpuweb#3338, by @kainino0x.
-
Expire the current texture with a microtask gpuweb/gpuweb#3337, by @toji.
-
Add a static alias analysis gpuweb/gpuweb#3336, by @alan-baker.
-
Update pull request preview action. gpuweb/gpuweb#3335, by @dj2.
-
Update remaining method algorithms to new timeline style gpuweb/gpuweb#3329, by @toji.
-
wgsl: Fix the description of inverse triangle functions gpuweb/gpuweb#3327, by @jiangzhaoming.
-
Add missing timeline headers in getMappedRange and unmap gpuweb/gpuweb#3325, by @kainino0x.
-
Convert Create* methods to new timeline style gpuweb/gpuweb#3322, by @toji.
-
Grammar: collect assignment, incr, decr into "variable_updating_statement" gpuweb/gpuweb#3321, by @dneto0.
-
uniformity: Clarify partial/full assignment fo gpuweb/gpuweb#3320, by @dneto0.
-
Reduce max binding index and make it a limit gpuweb/gpuweb#3318, by @toji.
-
Clarify default mixed in with cases, fix rule about duplicate case values gpuweb/gpuweb#3316, by @dneto0.
-
Remove fallthrough statements gpuweb/gpuweb#3315, by @alan-baker.
-
[editorial] Function chapter clarifications gpuweb/gpuweb#3314, by @alan-baker.
-
Update WGSL editors gpuweb/gpuweb#3313, by @dneto0.
-
Lower the maximum binding index in createBindGroupLayout to 640 gpuweb/gpuweb#3312, by @nical.
-
When source usability is bad, error asynchronously gpuweb/gpuweb#3311, by @kainino0x.
-
Wait for queue completion before destroying device gpuweb/gpuweb#3310, by @kainino0x.
-
[editorial] Replace uses of integral as a type gpuweb/gpuweb#3309, by @alan-baker.
-
[editorial] Add links in the atomic types section gpuweb/gpuweb#3308, by @alan-baker.
-
Clarify stores to invalid memory references gpuweb/gpuweb#3307, by @alan-baker.
-
Clarify that type aliases can be used for type constructors gpuweb/gpuweb#3306, by @dneto0.
-
Clarify limits on workgroup size dimension values gpuweb/gpuweb#3305, by @dneto0.
-
[editorial] Attribute clarifications gpuweb/gpuweb#3304, by @alan-baker.
-
[editorial] Fixes for declaration and scope gpuweb/gpuweb#3303, by @alan-baker.
-
Disallow duplicated built-in inputs and outputs gpuweb/gpuweb#3302, by @alan-baker.
-
[editorial] Fix markdown gpuweb/gpuweb#3301, by @alan-baker.
-
Separate content/device internal states of GPUBuffer gpuweb/gpuweb#3300, by @kainino0x.
-
Rework WGSL grammar to make it nicer for recursive descent gpuweb/gpuweb#3299, by @dneto0.
-
Add a data-flow analysis for local variables in uniformity gpuweb/gpuweb#3298, by @alan-baker.
-
publish grammar without left-recursion; add tooling to make it so gpuweb/gpuweb#3292, by @dneto0.
-
Add predeclared short names for vector types gpuweb/gpuweb#3291, by @dneto0.
-
Remove non-generic types from keywords list. gpuweb/gpuweb#3285, by @dj2.
-
Move non-type keywords into the other keyword section. gpuweb/gpuweb#3284, by @dj2.
-
Add external_texture to Resource Layout Compatibility table gpuweb/gpuweb#3283, by @dneto0.
-
wgsl: Fix typo in cosh(x) accuracy definition gpuweb/gpuweb#3282, by @zoddicus.
-
Remove types from keyword list. gpuweb/gpuweb#3278, by @dj2.
-
Clarify effect of @builtin struct member when not an entry point IO gpuweb/gpuweb#3275, by @dneto0.
-
analyze/lalr.py: Add -simple option to dump un-canonicalized grammar gpuweb/gpuweb#3274, by @dneto0.
-
-
83 pull requests merged.
-
Add note about texture copies not preserving bit values gpuweb/gpuweb#3397.
-
Remove 'invariant' from reserved word list gpuweb/gpuweb#3395.
-
Pull reserved word list into its own file gpuweb/gpuweb#3394.
-
Add the "internal" error type gpuweb/gpuweb#3123.
-
[editorial] Clarify specification of stripIndexFormat gpuweb/gpuweb#3396.
-
[editorial] consistent x-refs to IDL methods, add a common dfn for setBindGroup gpuweb/gpuweb#3398.
-
Errors for constant out-of-bounds accesses gpuweb/gpuweb#3400.
-
Make error reporting unordered gpuweb/gpuweb#3362.
-
Preview includes gpuweb/gpuweb#3393.
-
Clarify units for trig and hyperbolic functions gpuweb/gpuweb#3392.
-
[editorial] highlight data-link-type=idl too gpuweb/gpuweb#3391.
-
[editorial] Define scheme for state on different timelines gpuweb/gpuweb#3382.
-
Deploy previews to GitHub gpuweb/gpuweb#3384.
-
Remove device-timeline validation error in unmap() gpuweb/gpuweb#3368.
-
Require maxBufferSize to be at least as max*BufferBindingSize gpuweb/gpuweb#3380.
-
Fix declaration of [SameObject]-ness on interface rather than dictionary gpuweb/gpuweb#3385.
-
Fix address space for runtime-sized array example gpuweb/gpuweb#3388.
-
Fix PR previews with diagrams gpuweb/gpuweb#3387.
-
[editorial] Improve exposition of the uniformity analysis gpuweb/gpuweb#3374.
-
Reduce max binding index and make it a limit gpuweb/gpuweb#3318.
-
Move promise resolutions in algorithms to content timeline gpuweb/gpuweb#3378.
-
[editorial] Make `GPUUncapturedErrorEvent.error` `[SameObject]` gpuweb/gpuweb#3381.
-
[editorial] Vertical column headers for texture format table gpuweb/gpuweb#3369.
-
Move timestamp query inside passes into a separate extension doc gpuweb/gpuweb#3370.
-
Add timestampWrites to the section of Timestamp Query gpuweb/gpuweb#3371.
-
Style links with the timeline of the reference they point to gpuweb/gpuweb#3375.
-
[editorial] Specify what part of buffer is updated in unmap() gpuweb/gpuweb#3353.
-
Make getBGL return a "copy" instead of a "reference" gpuweb/gpuweb#3341.
-
Rename 'Creation' headers for disambiguation gpuweb/gpuweb#3373.
-
[editorial] Fix uniformity for statements table render gpuweb/gpuweb#3372.
-
Add per-pixel accounting of render target depths gpuweb/gpuweb#3367.
-
Add maxBufferSize limit gpuweb/gpuweb#3366.
-
When source usability is bad, error asynchronously gpuweb/gpuweb#3311.
-
Specify OOM behavior in mapping gpuweb/gpuweb#3351.
-
publish grammar without left-recursion; add tooling to make it so gpuweb/gpuweb#3292.
-
Define promise resolution ordering gpuweb/gpuweb#3361.
-
[editorial] Clarify 24-bit depth gpuweb/gpuweb#3360.
-
Require render and resolve to have same base format gpuweb/gpuweb#3363.
-
[editorial] Fix logo's gpuweb/gpuweb#3352.
-
Reject redundant map requests early so there are never multiple pending promises gpuweb/gpuweb#3348.
-
Better formalize and document mapAsync promise ordering gpuweb/gpuweb#3338.
-
Add tracking-vector annotations gpuweb/gpuweb#3340.
-
Update the WebGPU editors list gpuweb/gpuweb#3350.
-
[Editorial] Clarifying Abstract Types in the first line gpuweb/gpuweb#3343.
-
Expose the buffer state as an attribute gpuweb/gpuweb#3014.
-
Add proposals/ directory and pipeline statistics extension doc gpuweb/gpuweb#2301.
-
[editorial] Fix attribute typos gpuweb/gpuweb#3345.
-
[editorial] Introduction clarifications gpuweb/gpuweb#3344.
-
Update remaining method algorithms to new timeline style gpuweb/gpuweb#3329.
-
Set a limit for maximum array size gpuweb/gpuweb#3078.
-
Add missing timeline headers in getMappedRange and unmap gpuweb/gpuweb#3325.
-
Convert Create* methods to new timeline style gpuweb/gpuweb#3322.
-
wgsl: Fix the description of inverse triangle functions gpuweb/gpuweb#3327.
-
Grammar: collect assignment, incr, decr into "variable_updating_statement" gpuweb/gpuweb#3321.
-
uniformity: Clarify partial/full assignment fo gpuweb/gpuweb#3320.
-
Clarify default mixed in with cases, fix rule about duplicate case values gpuweb/gpuweb#3316.
-
Add a data-flow analysis for local variables in uniformity gpuweb/gpuweb#3298.
-
Add external_texture to Resource Layout Compatibility table gpuweb/gpuweb#3283.
-
Separate content/device internal states of GPUBuffer gpuweb/gpuweb#3300.
-
Wait for queue completion before destroying device gpuweb/gpuweb#3310.
-
[editorial] Add links in the atomic types section gpuweb/gpuweb#3308.
-
[editorial] Replace uses of integral as a type gpuweb/gpuweb#3309.
-
[editorial] Function chapter clarifications gpuweb/gpuweb#3314.
-
Remove fallthrough statements gpuweb/gpuweb#3315.
-
Update WGSL editors gpuweb/gpuweb#3313.
-
Clarify stores to invalid memory references gpuweb/gpuweb#3307.
-
Clarify that type aliases can be used for type constructors gpuweb/gpuweb#3306.
-
Clarify limits on workgroup size dimension values gpuweb/gpuweb#3305.
-
[editorial] Attribute clarifications gpuweb/gpuweb#3304.
-
Disallow duplicated built-in inputs and outputs gpuweb/gpuweb#3302.
-
Rework WGSL grammar to make it nicer for recursive descent gpuweb/gpuweb#3299.
-
[editorial] Fixes for declaration and scope gpuweb/gpuweb#3303.
-
[editorial] Fix markdown gpuweb/gpuweb#3301.
-
Adjust argumentdef style to fit better within the width of the spec gpuweb/gpuweb#3239.
-
Move non-type keywords into the other keyword section. gpuweb/gpuweb#3284.
-
Allow using depth textures with 'unfilterable-float' bindings gpuweb/gpuweb#3230.
-
wgsl: Fix typo in cosh(x) accuracy definition gpuweb/gpuweb#3282.
-
Clarify effect of @builtin struct member when not an entry point IO gpuweb/gpuweb#3275.
-
clarify lifetime of a declaration in a loop body gpuweb/gpuweb#3267.
-
Add @workgroup_size attributes on compute shader examples gpuweb/gpuweb#3269.
-
analyze/lalr.py: Add -simple option to dump un-canonicalized grammar gpuweb/gpuweb#3274.
-
Fix typo in example of user-defined function gpuweb/gpuweb#3268.
-
Fix repeated 'keyword' in explanation of extension names gpuweb/gpuweb#3270.
-
-
12 pull requests closed.
-
Add `multi-draw-indirect` feature. gpuweb/gpuweb#1949.
-
Queue kinds and discovery gpuweb/gpuweb#1306.
-
Add shader-float16 extension gpuweb/gpuweb#1275.
-
Add callback entrypoints for per-frame Promise entrypoints. gpuweb/gpuweb#626.
-
Multi-queue synchronization gpuweb/gpuweb#1169.
-
Multiple Queues skeleton proposal. gpuweb/gpuweb#95.
-
Allow e.g. `if (cond) break/continue/return/kill;` without brackets. gpuweb/gpuweb#705.
-
Update pull request preview action. gpuweb/gpuweb#3335.
-
Lower the maximum binding index in createBindGroupLayout to 640 gpuweb/gpuweb#3312.
-
Separate subsection for each method on GPUQueue gpuweb/gpuweb#3231.
-
Simplify device-timeline map states gpuweb/gpuweb#3264.
-
Remove types from keyword list. gpuweb/gpuweb#3278.
-
-
71 pull requests received 316 new comments.
-
publish grammar without left-recursion; add tooling to make it so gpuweb/gpuweb#3292 (18 comments by @dj2, @dneto0)
-
Move promise resolutions in algorithms to content timeline gpuweb/gpuweb#3378 (17 comments by @kainino0x, @toji)
-
Add a static alias analysis gpuweb/gpuweb#3336 (16 comments by @alan-baker, @djg, @dneto0, @jimblandy, @krogovin)
-
Add a data-flow analysis for local variables in uniformity gpuweb/gpuweb#3298 (15 comments by @alan-baker, @dneto0)
-
[editorial] Uniformity examples gpuweb/gpuweb#3379 (14 comments by @alan-baker, @dneto0, @kainino0x, @mehmetoguzderin)
-
Add per-pixel accounting of render target depths gpuweb/gpuweb#3367 (13 comments by @kainino0x, @litherum)
-
Adjust argumentdef style to fit better within the width of the spec gpuweb/gpuweb#3239 (9 comments by @kainino0x, @toji)
-
[editorial] Improve exposition of the uniformity analysis gpuweb/gpuweb#3374 (8 comments by @alan-baker, @dneto0, @mehmetoguzderin)
-
When source usability is bad, error asynchronously gpuweb/gpuweb#3311 (7 comments by @junov, @kainino0x)
-
Clarify effect of @builtin struct member when not an entry point IO gpuweb/gpuweb#3275 (6 comments by @alan-baker, @dneto0)
-
Style links with the timeline of the reference they point to gpuweb/gpuweb#3375 (5 comments by @kainino0x, @toji)
-
Clarify default mixed in with cases, fix rule about duplicate case values gpuweb/gpuweb#3316 (5 comments by @alan-baker, @dneto0)
-
Lower the maximum binding index in createBindGroupLayout to 640 gpuweb/gpuweb#3312 (5 comments by @kainino0x, @Kangz, @toji)
-
analyze/lalr.py: Add -simple option to dump un-canonicalized grammar gpuweb/gpuweb#3274 (5 comments by @dj2, @dneto0)
-
Pull reserved word list into its own file gpuweb/gpuweb#3394 (4 comments by @dneto0, @mehmetoguzderin)
-
Describe validation errors from createShaderModule gpuweb/gpuweb#3347 (4 comments by @dneto0, @kainino0x)
-
Make GPUShaderModuleCompilationHint.layout optional gpuweb/gpuweb#2876 (4 comments by @juj, @kainino0x, @toji)
-
Update pull request preview action. gpuweb/gpuweb#3335 (4 comments by @dj2, @kainino0x, @mehmetoguzderin)
-
Add predeclared short names for vector types gpuweb/gpuweb#3291 (4 comments by @alan-baker, @dneto0)
-
Rework WGSL grammar to make it nicer for recursive descent gpuweb/gpuweb#3299 (3 comments by @alan-baker, @dneto0)
-
Remove device-timeline validation error in unmap() gpuweb/gpuweb#3368 (2 comments by @kainino0x, @Kangz)
-
[editorial] Define scheme for state on different timelines gpuweb/gpuweb#3382 (2 comments by @kainino0x, @toji)
-
Reject redundant map requests early so there are never multiple pending promises gpuweb/gpuweb#3348 (2 comments by @kainino0x, @nical)
-
Add proposals/ directory and pipeline statistics extension doc gpuweb/gpuweb#2301 (2 comments by @kainino0x)
-
Update remaining method algorithms to new timeline style gpuweb/gpuweb#3329 (2 comments by @kainino0x)
-
Clarify limits on workgroup size dimension values gpuweb/gpuweb#3305 (2 comments by @alan-baker, @dneto0)
-
Separate subsection for each method on GPUQueue gpuweb/gpuweb#3231 (2 comments by @kainino0x, @toji)
-
Fix declaration of [SameObject]-ness on interface rather than dictionary gpuweb/gpuweb#3385 (1 comment by @kainino0x)
-
Add callback entrypoints for per-frame Promise entrypoints. gpuweb/gpuweb#626 (1 comment by @kainino0x)
-
[editorial] Specify what part of buffer is updated in unmap() gpuweb/gpuweb#3353 (1 comment by @kainino0x)
-
Expose different error types directly instead of via polymorphism gpuweb/gpuweb#2750 (1 comment by @juj)
-
@size attrib only applied to member types with creation-fixed footprint gpuweb/gpuweb#3157 (1 comment by @kdashg)
gpuweb/cts
-
25 issues opened.
-
wgsl: 'invariant' is no longer a reserved word gpuweb/cts#1817, by @dneto0.
-
wgsl: shader-creation and pipeline-creation errors for invalid expressions gpuweb/cts#1814, by @dneto0.
-
wgsl: Add idiomatic support for f16 values gpuweb/cts#1812, by @zoddicus.
-
wgsl: Cleanup FP helper functions for implicit f32-ness gpuweb/cts#1808, by @zoddicus.
-
Unify preview deployment pipeline with main repository gpuweb/cts#1804, by @mehmetoguzderin.
-
val: Handle missing validations in copyT2B and copyB2T gpuweb/cts#1799, by @Gyuyoung.
-
wgsl: Excise dead code from FP framework rewrite gpuweb/cts#1794, by @zoddicus.
-
wgsl: `ldexp` tests might now be incorrectly failing SwiftShader gpuweb/cts#1791, by @zoddicus.
-
Test and improve `fullI32Range` gpuweb/cts#1785, by @zoddicus.
-
fullF32Range values are not particularly interesting gpuweb/cts#1782, by @zoddicus.
-
op: requestDevice gpuweb/cts#1781, by @lokokung.
-
wgsl: Don't use set when generating flushed/non-flushed number lists gpuweb/cts#1778, by @zoddicus.
-
Rename `make*F32IntervalCase` functions to `make*ToF32IntervalCase` gpuweb/cts#1774, by @zoddicus.
-
wgsl: division tests should cover whole span of f32 gpuweb/cts#1764, by @zoddicus.
-
wgsl: f32 execution tests for `remainder (%)` gpuweb/cts#1763, by @zoddicus.
-
Optimize input data generation gpuweb/cts#1757, by @austinEng.
-
Track all the buffers and textures with TrackForCleanup() gpuweb/cts#1752, by @Jiawei-Shao.
-
wgsl: 'fallthrough' removed, and 'default' can be one of the case values gpuweb/cts#1750, by @dneto0.
-
Make FP testing framework (f32_interval.ts) more idiomatic gpuweb/cts#1747, by @zoddicus.
-
wgsl: disallow duplicated pipeline builtin inputs, and outputs gpuweb/cts#1745, by @dneto0.
-
val: Test beginComputePass and GPUComputePassDescriptor validation gpuweb/cts#1743, by @Gyuyoung.
-
Update val:CopyExternalImageToTexture:OOB,source gpuweb/cts#1738, by @Gyuyoung.
-
-
14 issues closed.
-
Test and improve `fullI32Range` gpuweb/cts#1785.
-
wgsl: Excise dead code from FP framework rewrite gpuweb/cts#1794.
-
op: requestDevice gpuweb/cts#1781.
-
val: capability_checks,features errors should be exceptions gpuweb/cts#919.
-
val: resource_usages,texture,in_pass_encoder gpuweb/cts#923.
-
fullF32Range values are not particularly interesting gpuweb/cts#1782.
-
Rename `make*F32IntervalCase` functions to `make*ToF32IntervalCase` gpuweb/cts#1774.
-
wgsl: division tests should cover whole span of f32 gpuweb/cts#1764.
-
wgsl: Don't use set when generating flushed/non-flushed number lists gpuweb/cts#1778.
-
Make FP testing framework (f32_interval.ts) more idiomatic gpuweb/cts#1747.
-
val: state,device_lost gpuweb/cts#882.
-
-
13 issues received 23 new comments.
-
wgsl: Add idiomatic support for f16 values gpuweb/cts#1812 (5 comments by @kainino0x, @tidoust, @zoddicus)
-
wgsl: `ldexp` tests might now be incorrectly failing SwiftShader gpuweb/cts#1791 (3 comments by @dneto0, @zoddicus)
-
val: capability_checks,features errors should be exceptions gpuweb/cts#919 (2 comments by @Gyuyoung, @kainino0x)
-
val: Test beginComputePass and GPUComputePassDescriptor validation gpuweb/cts#1743 (1 comment by @Gyuyoung)
-
-
61 pull requests opened.
-
Make `make*ToF32IntervalCase`s more idiomatic gpuweb/cts#1816, by @zoddicus.
-
Add `length` F32 tests gpuweb/cts#1815, by @zoddicus.
-
Add storage_texture,usage test to createVindGroup.spec.ts gpuweb/cts#1813, by @Gyuyoung.
-
Add `smoothstep` F32 Tests gpuweb/cts#1811, by @zoddicus.
-
Add 'multisampled_validation' test to createBindGroup.spec.ts gpuweb/cts#1810, by @Gyuyoung.
-
Use trackForCleanup() for dst texture in copyToTexture cases gpuweb/cts#1809, by @shaoboyan.
-
Implement 'texture_buffer_usage' test to buffer_texture_copies.spec.ts gpuweb/cts#1807, by @Gyuyoung.
-
Add RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/cts#1806, by @Jiawei-Shao.
-
Add 'texture_buffer_usage' test to buffer_texture_copies.spec.ts gpuweb/cts#1803, by @Gyuyoung.
-
Mark root .eslintrc.json as "root: true" gpuweb/cts#1802, by @austinEng.
-
Add global test config gpuweb/cts#1801, by @austinEng.
-
Add sample_count test to buffer_texture_copies.spec.ts gpuweb/cts#1800, by @Gyuyoung.
-
Adds remaining adapter.requestDevice tests. gpuweb/cts#1798, by @lokokung.
-
Fix linter warning in CI gpuweb/cts#1797, by @zoddicus.
-
Remove old implementation of correctlyRounded gpuweb/cts#1796, by @zoddicus.
-
Clarify comments in ULP code gpuweb/cts#1795, by @zoddicus.
-
Implement 'texture,device_mismatch' test in writeTexture.spec.ts gpuweb/cts#1793, by @Gyuyoung.
-
Add tests to `fullI32Range` gpuweb/cts#1792, by @zoddicus.
-
Add 'maximum_binding_limit' test to createBindGroupLayout.spec.ts gpuweb/cts#1790, by @Gyuyoung.
-
Use let declarations for literal input in shader,execution,expression gpuweb/cts#1788, by @jiangzhaoming.
-
Input data generation optimization gpuweb/cts#1787, by @toji.
-
Linearly spread `f32FullRange` values over possible f32 values gpuweb/cts#1786, by @zoddicus.
-
Check shader validation before testing results (alt version) gpuweb/cts#1784, by @kainino0x.
-
Refactor binary arithmetic tests to use common test values gpuweb/cts#1783, by @zoddicus.
-
webgpu/shader/execution: Check validation before testing results gpuweb/cts#1780, by @ben-clayton.
-
Remove pushing lists of numbers through Set to guarantee uniqueness gpuweb/cts#1779, by @zoddicus.
-
Update 'multisampled_validation' test in createBindGroupLayout.spec.ts gpuweb/cts#1777, by @Gyuyoung.
-
Add storage_texture,formats test to createBindGroupLayout.spec.ts gpuweb/cts#1776, by @Gyuyoung.
-
Rename `make*F32IntervalCase` to `make*ToF32IntervalCase` gpuweb/cts#1775, by @zoddicus.
-
Make F32Interval unittests more idiomatic gpuweb/cts#1773, by @zoddicus.
-
Add storage_texture,layout_dimension test to createBindGroupLayout.spec.ts gpuweb/cts#1772, by @Gyuyoung.
-
Add brackets to robust buffer access test. gpuweb/cts#1771, by @dj2.
-
Fix f16 and u32 testcases in webgpu:shader,validation,parse,literal:* gpuweb/cts#1769, by @jiangzhaoming.
-
Change .origin in GPUImageCopyExternalImage from 3D to 2D gpuweb/cts#1768, by @shaoboyan.
-
Updates error scope tests and remove outdated README. gpuweb/cts#1766, by @lokokung.
-
Implement `remainder (%)` f32 test gpuweb/cts#1765, by @zoddicus.
-
Add timestamp_query_set,device_mismatch test to beginRenderPass tests gpuweb/cts#1761, by @Gyuyoung.
-
Add timestamp_query_set,device_mismatch test to beginComputePass tests gpuweb/cts#1760, by @Gyuyoung.
-
Add _mimeType for input videos in external_texture cases gpuweb/cts#1759, by @shaoboyan.
-
Pass a dummy pipeline layout to createNoOpComputePipeline gpuweb/cts#1758, by @Gyuyoung.
-
Add indirect_dispatch_buffer,usage validation test gpuweb/cts#1756, by @Gyuyoung.
-
Refactor *SubnormalNumber to not depend on *SubnormalScalar gpuweb/cts#1755, by @zoddicus.
-
Use common `sparseF32Range` instead of defining it multiple different places gpuweb/cts#1754, by @zoddicus.
-
Track all the resources with trackForCleanup in copy and resource ini… gpuweb/cts#1753, by @Jiawei-Shao.
-
Add a query_set_type test to beginComputePass tests gpuweb/cts#1751, by @Gyuyoung.
-
Implementing `dot` f32 tests gpuweb/cts#1748, by @zoddicus.
-
Implement `asinh` f32 tests gpuweb/cts#1746, by @zoddicus.
-
Add beginComputePass validation test gpuweb/cts#1744, by @Gyuyoung.
-
Add more texture usages to image_copy,texture_related:usage test gpuweb/cts#1742, by @Gyuyoung.
-
Implement `atanh` f32 tests gpuweb/cts#1741, by @zoddicus.
-
Use containers for alternate interval generators gpuweb/cts#1740, by @zoddicus.
-
Update val:CopyExternalImageToTexture:OOB,source gpuweb/cts#1739, by @Gyuyoung.
-
Implement `acosh` f32 tests gpuweb/cts#1737, by @zoddicus.
-
Implement `tanh` f32 tests gpuweb/cts#1736, by @zoddicus.
-
Implement `cosh` f32 tests gpuweb/cts#1735, by @zoddicus.
-
Implement `sinh` f32 tests gpuweb/cts#1734, by @zoddicus.
-
Add 'VideoFrame' in 'SourceType' for external_texture cts gpuweb/cts#1733, by @shaoboyan.
-
Implement `sqrt` f32 tests gpuweb/cts#1732, by @zoddicus.
-
-
58 pull requests merged.
-
Add `smoothstep` F32 Tests gpuweb/cts#1811.
-
Add tests to `fullI32Range` gpuweb/cts#1792.
-
Remove old implementation of correctlyRounded gpuweb/cts#1796.
-
Add 'multisampled_validation' test to createBindGroup.spec.ts gpuweb/cts#1810.
-
Use trackForCleanup() for dst texture in copyToTexture cases gpuweb/cts#1809.
-
Adds remaining adapter.requestDevice tests. gpuweb/cts#1798.
-
Add RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/cts#1806.
-
Add 'texture_buffer_usage' test to buffer_texture_copies.spec.ts gpuweb/cts#1803.
-
Add sample_count test to buffer_texture_copies.spec.ts gpuweb/cts#1800.
-
Implement 'texture,device_mismatch' test in writeTexture.spec.ts gpuweb/cts#1793.
-
Add 'maximum_binding_limit' test to createBindGroupLayout.spec.ts gpuweb/cts#1790.
-
Add global test config gpuweb/cts#1801.
-
Mark root .eslintrc.json as "root: true" gpuweb/cts#1802.
-
Input data generation optimization gpuweb/cts#1787.
-
Clarify comments in ULP code gpuweb/cts#1795.
-
webgpu/shader/execution: Check validation before testing results gpuweb/cts#1780.
-
Linearly spread `f32FullRange` values over possible f32 values gpuweb/cts#1786.
-
Implement `remainder (%)` f32 test gpuweb/cts#1765.
-
Rename `make*F32IntervalCase` to `make*ToF32IntervalCase` gpuweb/cts#1775.
-
Add storage_texture,formats test to createBindGroupLayout.spec.ts gpuweb/cts#1776.
-
Refactor binary arithmetic tests to use common test values gpuweb/cts#1783.
-
Remove pushing lists of numbers through Set to guarantee uniqueness gpuweb/cts#1779.
-
wgsl: Replace module-scope 'let' with 'const' gpuweb/cts#1721.
-
Make F32Interval unittests more idiomatic gpuweb/cts#1773.
-
Updates error scope tests and remove outdated README. gpuweb/cts#1766.
-
Add brackets to robust buffer access test. gpuweb/cts#1771.
-
Change .origin in GPUImageCopyExternalImage from 3D to 2D gpuweb/cts#1768.
-
Add _mimeType for input videos in external_texture cases gpuweb/cts#1759.
-
Refactor *SubnormalNumber to not depend on *SubnormalScalar gpuweb/cts#1755.
-
Implementing `dot` f32 tests gpuweb/cts#1748.
-
Add indirect_dispatch_buffer,usage validation test gpuweb/cts#1756.
-
Update compositingAlphaMode -> alphaMode gpuweb/cts#1667.
-
Add a query_set_type test to beginComputePass tests gpuweb/cts#1751.
-
Add more texture usages to image_copy,texture_related:usage test gpuweb/cts#1742.
-
Add beginComputePass validation test gpuweb/cts#1744.
-
Implement `asinh` f32 tests gpuweb/cts#1746.
-
Implement `atanh` f32 tests gpuweb/cts#1741.
-
Implement `acosh` f32 tests gpuweb/cts#1737.
-
Implement `tanh` f32 tests gpuweb/cts#1736.
-
Use containers for alternate interval generators gpuweb/cts#1740.
-
Update val:CopyExternalImageToTexture:OOB,source gpuweb/cts#1739.
-
upgrade typescript to 4.7.4 gpuweb/cts#1713.
-
Implement GPURenderPassDescriptor.maxDrawCount test gpuweb/cts#1693.
-
Add 'VideoFrame' in 'SourceType' for external_texture cts gpuweb/cts#1733.
-
Implement `cosh` f32 tests gpuweb/cts#1735.
-
Implement `sinh` f32 tests gpuweb/cts#1734.
-
Implement `sqrt` f32 tests gpuweb/cts#1732.
-
-
3 pull requests closed.
-
Fix linter warning in CI gpuweb/cts#1797.
-
Check shader validation before testing results (alt version) gpuweb/cts#1784.
-
Pass a dummy pipeline layout to createNoOpComputePipeline gpuweb/cts#1758.
-
-
2 pull requests reopened.
-
Add sample_count test to buffer_texture_copies.spec.ts gpuweb/cts#1800, by @Gyuyoung.
-
Add indirect_dispatch_buffer,usage validation test gpuweb/cts#1756, by @Gyuyoung.
-
-
43 pull requests received 202 new comments.
-
Updates error scope tests and remove outdated README. gpuweb/cts#1766 (12 comments by @kainino0x, @lokokung)
-
Change .origin in GPUImageCopyExternalImage from 3D to 2D gpuweb/cts#1768 (10 comments by @Gyuyoung, @kainino0x, @shaoboyan)
-
Add _mimeType for input videos in external_texture cases gpuweb/cts#1759 (10 comments by @kainino0x, @shaoboyan)
-
Add more texture usages to image_copy,texture_related:usage test gpuweb/cts#1742 (8 comments by @Gyuyoung, @kainino0x)
-
Update 'multisampled_validation' test in createBindGroupLayout.spec.ts gpuweb/cts#1777 (7 comments by @Gyuyoung, @kainino0x)
-
Use common `sparseF32Range` instead of defining it multiple different places gpuweb/cts#1754 (6 comments by @ben-clayton, @zoddicus)
-
Pass a dummy pipeline layout to createNoOpComputePipeline gpuweb/cts#1758 (6 comments by @Gyuyoung, @kainino0x)
-
Add sample_count test to buffer_texture_copies.spec.ts gpuweb/cts#1800 (5 comments by @Gyuyoung, @kainino0x)
-
webgpu/shader/execution: Check validation before testing results gpuweb/cts#1780 (5 comments by @ben-clayton, @kainino0x)
-
Add 'texture_buffer_usage' test to buffer_texture_copies.spec.ts gpuweb/cts#1803 (4 comments by @Gyuyoung, @kainino0x)
-
Add visibility,VERTEX_shader_stage_storage_texture_access test to createBindGroupLayout.spec.ts gpuweb/cts#1770 (4 comments by @Gyuyoung, @kainino0x)
-
Update val:CopyExternalImageToTexture:OOB,source gpuweb/cts#1739 (4 comments by @Gyuyoung, @kainino0x, @shaoboyan)
-
Add 'maximum_binding_limit' test to createBindGroupLayout.spec.ts gpuweb/cts#1790 (3 comments by @Gyuyoung, @kainino0x)
-
Implement 'texture_buffer_usage' test to buffer_texture_copies.spec.ts gpuweb/cts#1807 (2 comments by @Gyuyoung, @kainino0x)
-
Use let declarations for literal input in shader,execution,expression gpuweb/cts#1788 (2 comments by @ben-clayton, @jiangzhaoming)
-
Remove pushing lists of numbers through Set to guarantee uniqueness gpuweb/cts#1779 (2 comments by @ben-clayton, @zoddicus)
-
Add a query_set_type test to beginComputePass tests gpuweb/cts#1751 (2 comments by @Gyuyoung, @kainino0x)
-
Implement 'texture,device_mismatch' test in writeTexture.spec.ts gpuweb/cts#1793 (1 comment by @Gyuyoung)
-
Add storage_texture,formats test to createBindGroupLayout.spec.ts gpuweb/cts#1776 (1 comment by @kainino0x)
-
Add storage_texture,layout_dimension test to createBindGroupLayout.spec.ts gpuweb/cts#1772 (1 comment by @Gyuyoung)
-
Fix f16 and u32 testcases in webgpu:shader,validation,parse,literal:* gpuweb/cts#1769 (1 comment by @Jiawei-Shao)
-
Track all the resources with trackForCleanup in copy and resource ini… gpuweb/cts#1753 (1 comment by @Jiawei-Shao)
TR publications
-
21 documents published.
-
WebGPU published 6 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 15 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/cts
-
66 issues opened.
-
wgsl: enforce that declared array element count is element_count_expression gpuweb/cts#1731, by @dneto0.
-
op: device,destroy gpuweb/cts#1730, by @kainino0x.
-
wgsl: uniformity update: function returns are a reconvergence point gpuweb/cts#1720, by @dneto0.
-
wgsl: some @attrib parameters can be expressions gpuweb/cts#1716, by @dneto0.
-
wgsl: @size attrib only applies to struct members with creation-fixed footprint gpuweb/cts#1715, by @dneto0.
-
wgsl: rename staticAssert to static_assert gpuweb/cts#1714, by @dneto0.
-
wsgl: disallow repeated use of builtin inputs and builtin outputs gpuweb/cts#1712, by @dneto0.
-
wgsl: Validate limits gpuweb/cts#1710, by @dj2.
-
wgsl: Validate program gpuweb/cts#1709, by @dj2.
-
wgsl: validate `enable` gpuweb/cts#1708, by @dj2.
-
wgsl: value `(@group, @binding)` pair must be unique gpuweb/cts#1707, by @dj2.
-
wgsl: Validate that `@location` can duplicate between input/output gpuweb/cts#1706, by @dj2.
-
wgsl: support abstract arrays, and constructor that infers result type. gpuweb/cts#1705, by @dneto0.
-
Add tests for dynamic index of abstract numeric composite gpuweb/cts#1698, by @ben-clayton.
-
op: buffer/device.destroy() unmaps buffer on current thread gpuweb/cts#1697, by @shrekshao.
-
wgsl: interpolation execution tests gpuweb/cts#1696, by @dj2.
-
val: Test copyExternalImageToTexture with dimensions gpuweb/cts#1694, by @Gyuyoung.
-
val: writeTexture validation that isn't common with copyB2T/T2B gpuweb/cts#1692, by @kainino0x.
-
wgsl: validate workgroup_size gpuweb/cts#1691, by @dj2.
-
wgsl: validate function restrictions gpuweb/cts#1690, by @dj2.
-
wgsl: validate creation-time functions gpuweb/cts#1689, by @dj2.
-
wgsl : validate function call parameter evaluation ordering gpuweb/cts#1688, by @dj2.
-
wgsl: validate user-defined functions gpuweb/cts#1687, by @dj2.
-
wgsl: validate `static_assert` statement gpuweb/cts#1686, by @dj2.
-
wgsl: validate `;` a valid statement gpuweb/cts#1685, by @dj2.
-
wgsl: validate `function_call` statement gpuweb/cts#1684, by @dj2.
-
wgsl: Execution tests for `discard` statement gpuweb/cts#1683, by @dj2.
-
wgsl: validate `return` statement gpuweb/cts#1682, by @dj2.
-
wgsl: validate `continue` statement gpuweb/cts#1681, by @dj2.
-
`x` & `y` are swapped in `atan2Interval` unittests gpuweb/cts#1680, by @zoddicus.
-
`atan2` tests are incorrectly skipping some combinations gpuweb/cts#1678, by @zoddicus.
-
wgsl: validate `break-if` statement gpuweb/cts#1672, by @dj2.
-
wgsl: validate `break` statement gpuweb/cts#1671, by @dj2.
-
wgsl: validate `while` statement gpuweb/cts#1670, by @dj2.
-
wgsl: validate `for` statement gpuweb/cts#1669, by @dj2.
-
wgsl: validate `loop` statement gpuweb/cts#1666, by @dj2.
-
val: Enhance `ranges` test in writeBuffer.spec.ts gpuweb/cts#1662, by @Gyuyoung.
-
wgsl: validate switch statements gpuweb/cts#1648, by @dj2.
-
wgsl: validate if statements gpuweb/cts#1647, by @dj2.
-
wgsl: Validate increment and decrement gpuweb/cts#1646, by @dj2.
-
wgsl: validate assignment gpuweb/cts#1644, by @dj2.
-
Handle infinite results of operations being potentially undefined gpuweb/cts#1642, by @zoddicus.
-
Support multiple discrete intervals being returned in FP testing framework gpuweb/cts#1641, by @zoddicus.
-
wgsl: user declarations can shadow predeclared objects gpuweb/cts#1636, by @dneto0.
-
wgsl: test saturate floating point builtin gpuweb/cts#1635, by @dneto0.
-
wgsl: compound statement tests gpuweb/cts#1634, by @dj2.
-
wgsl: indirection expression gpuweb/cts#1633, by @dj2.
-
wgsl: Address of test gpuweb/cts#1632, by @dj2.
-
wgsl: Bitwise tests gpuweb/cts#1631, by @dj2.
-
wgsl: Comparison expression tests gpuweb/cts#1627, by @dj2.
-
wgsl: Arithmetic Expression tests gpuweb/cts#1626, by @dj2.
-
CTS Interval implementations should not be arrow functions in builders gpuweb/cts#1625, by @zoddicus.
-
wgsl: Validate logical expressions gpuweb/cts#1623, by @dj2.
-
wgsl: validate struct accessor gpuweb/cts#1622, by @dj2.
-
wgsl: validate array access expressions gpuweb/cts#1621, by @dj2.
-
val: render pass descriptor gpuweb/cts#1618, by @Gyuyoung.
-
val: GPURenderPassDescriptor.maxDrawCount test gpuweb/cts#1613, by @takahirox.
-
wgsl: validate matrix access expressions gpuweb/cts#1611, by @dj2.
-
wgsl: validation vector access expressions gpuweb/cts#1610, by @dj2.
-
wgsl: execution test, bitcast: f16 cases remain gpuweb/cts#1609, by @dj2.
-
wgsl: Execution tests, constructors gpuweb/cts#1606, by @dj2.
-
wgsl: Parenthesized expressions gpuweb/cts#1605, by @dj2.
-
wgsl: Validate variable and value declaration parsing gpuweb/cts#1604, by @dj2.
-
wgsl: validation function scoped variables gpuweb/cts#1603, by @dj2.
-
-
17 issues closed.
-
wgsl: Shader does not require an entry point gpuweb/cts#1428.
-
wgsl: validate `@builtin` parsing gpuweb/cts#1446.
-
wgsl: Validate each `@builtin` value can be used as variable names gpuweb/cts#1443.
-
wgsl: Add builtin validation test for `frag_depth` gpuweb/cts#1437.
-
val: Enhance `ranges` test in writeBuffer.spec.ts gpuweb/cts#1662.
-
`atan2` tests are incorrectly skipping some combinations gpuweb/cts#1678.
-
`x` & `y` are swapped in `atan2Interval` unittests gpuweb/cts#1680.
-
How should accuracy of floating point builtins be evaluated? gpuweb/cts#792.
-
Handle infinite results of operations being potentially undefined gpuweb/cts#1642.
-
val: createComputePipeline gpuweb/cts#1555.
-
val: createRenderPipeline gpuweb/cts#1553.
-
wgsl: Test line break characters gpuweb/cts#1430.
-
-
10 issues received 13 new comments.
-
wgsl: Validate `@builtin` on non-entrypoint parameter, non-entrypoint return type or struct member fails gpuweb/cts#1435 (2 comments by @dj2)
-
op: buffer/device.destroy() unmaps buffer on current thread gpuweb/cts#1697 (2 comments by @kainino0x, @shrekshao)
-
Add tests for dynamic index of abstract numeric composite gpuweb/cts#1698 (1 comment by @ben-clayton)
-
68 pull requests opened.
-
Validate a shader without an entry point is valid. gpuweb/cts#1729, by @dj2.
-
Add non-entrypoint tests to the builtin validation. gpuweb/cts#1728, by @dj2.
-
Trailing comma in `@align` should be valid. gpuweb/cts#1727, by @dj2.
-
Add test to validate parsing of `@builtin` gpuweb/cts#1726, by @dj2.
-
Add test to allow builtin values as identifiers gpuweb/cts#1725, by @dj2.
-
Implement `trunc` f32 tests gpuweb/cts#1724, by @zoddicus.
-
Add `frag_depth` builtin validation. gpuweb/cts#1723, by @dj2.
-
shader/execution: Optimize "const" shader exeuction gpuweb/cts#1722, by @ben-clayton.
-
wgsl: Replace module-scope 'let' with 'const' gpuweb/cts#1721, by @ben-clayton.
-
Implement `mix` f32 tests gpuweb/cts#1719, by @zoddicus.
-
Implement `round` f32 tests gpuweb/cts#1718, by @zoddicus.
-
Implement `step` f32 tests gpuweb/cts#1717, by @zoddicus.
-
upgrade typescript to 4.7.4 gpuweb/cts#1713, by @kainino0x.
-
Add WebCodec VideoFrame as input source for GPUExternalTexture cases gpuweb/cts#1711, by @shaoboyan.
-
Fix testing of the offset in image_copy,layout_related:required_bytes… gpuweb/cts#1704, by @Kangz.
-
Implement val:queue,writeTexture:sample_count gpuweb/cts#1703, by @Gyuyoung.
-
Implement val:queue,writeTexture:texture_state gpuweb/cts#1702, by @Gyuyoung.
-
Implement `sign` f32 tests gpuweb/cts#1701, by @zoddicus.
-
Implement `saturate` f32 tests gpuweb/cts#1700, by @zoddicus.
-
Expand map_detach test for buffer should detach on device.destroy gpuweb/cts#1699, by @shrekshao.
-
Implement val:CopyExternalImageToTexture,destination_texture,dimension gpuweb/cts#1695, by @Gyuyoung.
-
Implement GPURenderPassDescriptor.maxDrawCount test gpuweb/cts#1693, by @takahirox.
-
Generate full range of `atan2` cases gpuweb/cts#1679, by @zoddicus.
-
Correct the order of params in `atan2` logging message gpuweb/cts#1677, by @zoddicus.
-
Implement `radians` f32 tests gpuweb/cts#1676, by @zoddicus.
-
Implement val:queue:writeTexture:usage gpuweb/cts#1675, by @Gyuyoung.
-
Enhance val:render_pass_descriptor:depth_stencil_attachment_depth_clear_value gpuweb/cts#1674, by @Gyuyoung.
-
Expand readbackFromWebGPUCanvas test with alphaMode gpuweb/cts#1673, by @shrekshao.
-
Implement f32 `degrees` tests gpuweb/cts#1668, by @zoddicus.
-
Update compositingAlphaMode -> alphaMode gpuweb/cts#1667, by @shrekshao.
-
Remove obsolete f32 testing infrastructure gpuweb/cts#1665, by @zoddicus.
-
Add validation tests for `@align`. gpuweb/cts#1664, by @dj2.
-
Add an out-of-range check to the ranges test in writeBuffer.spec.ts gpuweb/cts#1663, by @Gyuyoung.
-
Submit all shader batches before checking for results gpuweb/cts#1661, by @ben-clayton.
-
Make `render_pass_descriptor:resolveTarget,usage` test a valid case together gpuweb/cts#1660, by @Gyuyoung.
-
Convert `ldexp` tests to use interval framework gpuweb/cts#1659, by @zoddicus.
-
Add f32 tests for `pow` gpuweb/cts#1658, by @zoddicus.
-
Add reftests: canvas use as image source with opaque/premultiplied alpha mode gpuweb/cts#1657, by @shrekshao.
-
Implement val:render_pass_descriptor:timestampWrite,same_query_index gpuweb/cts#1656, by @Gyuyoung.
-
Update f32 tests to handle returning undefined for infinity gpuweb/cts#1655, by @zoddicus.
-
Implement val:render_pass_descriptor:timestampWrite,query_index gpuweb/cts#1654, by @Gyuyoung.
-
Implement val:render_pass_descriptor:timestamp_writes_query_set_type gpuweb/cts#1653, by @Gyuyoung.
-
Implement val:render_pass_descriptor:occlusion_query_set_type gpuweb/cts#1652, by @Gyuyoung.
-
Succinct the test names of render_pass_descriptor.spec.ts (3/3) gpuweb/cts#1651, by @Gyuyoung.
-
Improve resource_usages,texture,in_pass_encoder tests gpuweb/cts#1650, by @kainino0x.
-
Add context configuration tests gpuweb/cts#1649, by @toji.
-
Convert `clamp` f32 tests to use interval framework gpuweb/cts#1645, by @zoddicus.
-
Add support for multiple intervals to FP testing framework gpuweb/cts#1643, by @zoddicus.
-
Implement val:render_pass_descriptor:depth_stencil_attachment_depth_clear_value gpuweb/cts#1640, by @Gyuyoung.
-
Succinct the test names of render_pass_descriptor.spec.ts (2/n) gpuweb/cts#1639, by @Gyuyoung.
-
Add tests for getCurrentTexture and getPreferredCanvasFormat gpuweb/cts#1638, by @toji.
-
Generate many WPT html files for infra that doesn't support variants gpuweb/cts#1637, by @kainino0x.
-
Implement val:render_pass_descriptor:timestamp_writes_location gpuweb/cts#1630, by @Gyuyoung.
-
Correct logInterval & log2Interval handling of out of bounds inputs gpuweb/cts#1629, by @zoddicus.
-
Separate out interval operation definitions from builder functions gpuweb/cts#1628, by @zoddicus.
-
Implement val:render_pass_descriptor:occlusion_query_set_type gpuweb/cts#1620, by @Gyuyoung.
-
Succinct the test names of render_pass_descriptor.spec.ts (1/n) gpuweb/cts#1619, by @Gyuyoung.
-
Add test for required entry_point in shader. gpuweb/cts#1617, by @dj2.
-
Test line comments at EOF. gpuweb/cts#1616, by @dj2.
-
Rename entry_point test file. gpuweb/cts#1615, by @dj2.
-
Add GPURenderPassDescriptor.maxDrawCount test plan gpuweb/cts#1614, by @takahirox.
-
Convert `fract` tests to use interval framework gpuweb/cts#1612, by @zoddicus.
-
Minor clean up in life_of.md gpuweb/cts#1608, by @takahirox.
-
Add f32 tests for `tan` gpuweb/cts#1607, by @zoddicus.
-
Shorten the vertex buffer OOB validation test name gpuweb/cts#1602, by @jiangzhaoming.
-
Guard against writeTimestamp() calls without the proper feature gpuweb/cts#1601, by @toji.
-
Test zero levels or layers is invalid in render_pass,resolve:resolve_attachment gpuweb/cts#1600, by @Gyuyoung.
-
Update the descriptions of the tests of in_pass_encoder.spec.ts gpuweb/cts#1598, by @Gyuyoung.
-
-
64 pull requests merged.
-
Expand readbackFromWebGPUCanvas test with alphaMode gpuweb/cts#1673.
-
Implement `mix` f32 tests gpuweb/cts#1719.
-
Validate a shader without an entry point is valid. gpuweb/cts#1729.
-
Add non-entrypoint tests to the builtin validation. gpuweb/cts#1728.
-
Add test to validate parsing of `@builtin` gpuweb/cts#1726.
-
Trailing comma in `@align` should be valid. gpuweb/cts#1727.
-
Add test to allow builtin values as identifiers gpuweb/cts#1725.
-
Implement `trunc` f32 tests gpuweb/cts#1724.
-
Add `frag_depth` builtin validation. gpuweb/cts#1723.
-
shader/execution: Optimize "const" shader exeuction gpuweb/cts#1722.
-
Implement `round` f32 tests gpuweb/cts#1718.
-
Implement `step` f32 tests gpuweb/cts#1717.
-
Implement `sign` f32 tests gpuweb/cts#1701.
-
Implement `saturate` f32 tests gpuweb/cts#1700.
-
Implement `radians` f32 tests gpuweb/cts#1676.
-
Add WebCodec VideoFrame as input source for GPUExternalTexture cases gpuweb/cts#1711.
-
Add validation tests for `@align`. gpuweb/cts#1664.
-
Implement val:queue,writeTexture:sample_count gpuweb/cts#1703.
-
Implement val:queue,writeTexture:texture_state gpuweb/cts#1702.
-
Expand map_detach test for buffer should detach on device.destroy gpuweb/cts#1699.
-
Add an out-of-range check to the ranges test in writeBuffer.spec.ts gpuweb/cts#1663.
-
Implement val:queue:writeTexture:usage gpuweb/cts#1675.
-
Add GPURenderPassDescriptor.maxDrawCount test plan gpuweb/cts#1614.
-
Generate full range of `atan2` cases gpuweb/cts#1679.
-
Correct the order of params in `atan2` logging message gpuweb/cts#1677.
-
Implement f32 `degrees` tests gpuweb/cts#1668.
-
Remove obsolete f32 testing infrastructure gpuweb/cts#1665.
-
Add f32 tests for `pow` gpuweb/cts#1658.
-
Submit all shader batches before checking for results gpuweb/cts#1661.
-
Convert `ldexp` tests to use interval framework gpuweb/cts#1659.
-
Update f32 tests to handle returning undefined for infinity gpuweb/cts#1655.
-
Implement val:render_pass_descriptor:timestampWrite,same_query_index gpuweb/cts#1656.
-
Implement val:render_pass_descriptor:timestamp_writes_query_set_type gpuweb/cts#1653.
-
Implement val:render_pass_descriptor:timestampWrite,query_index gpuweb/cts#1654.
-
Implement val:render_pass_descriptor:occlusion_query_set_type gpuweb/cts#1652.
-
Succinct the test names of render_pass_descriptor.spec.ts (3/3) gpuweb/cts#1651.
-
Convert `clamp` f32 tests to use interval framework gpuweb/cts#1645.
-
Add support for multiple intervals to FP testing framework gpuweb/cts#1643.
-
Add context configuration tests gpuweb/cts#1649.
-
Add tests for getCurrentTexture and getPreferredCanvasFormat gpuweb/cts#1638.
-
Succinct the test names of render_pass_descriptor.spec.ts (2/n) gpuweb/cts#1639.
-
Implement val:render_pass_descriptor:timestamp_writes_location gpuweb/cts#1630.
-
Improve resource_usages,texture,in_pass_encoder tests gpuweb/cts#1650.
-
Succinct the test names of render_pass_descriptor.spec.ts (1/n) gpuweb/cts#1619.
-
Convert `fract` tests to use interval framework gpuweb/cts#1612.
-
Correct logInterval & log2Interval handling of out of bounds inputs gpuweb/cts#1629.
-
Separate out interval operation definitions from builder functions gpuweb/cts#1628.
-
Test line comments at EOF. gpuweb/cts#1616.
-
Rename entry_point test file. gpuweb/cts#1615.
-
Shorten the vertex buffer OOB validation test name gpuweb/cts#1602.
-
Add f32 tests for `tan` gpuweb/cts#1607.
-
Convert `inverseSqrt` tests to use interval framework gpuweb/cts#1597.
-
Convert subtraction tests to use interval framework gpuweb/cts#1594.
-
Update the descriptions of the tests of in_pass_encoder.spec.ts gpuweb/cts#1598.
-
Optimized several expensive methods in texture tests gpuweb/cts#1583.
-
Guard against writeTimestamp() calls without the proper feature gpuweb/cts#1601.
-
-
4 pull requests closed.
-
Add test for required entry_point in shader. gpuweb/cts#1617.
-
DNS: Prototype for how to implement inherited FP accuracy tests gpuweb/cts#1402.
-
Implement val:render_pass_descriptor:occlusion_query_set_type gpuweb/cts#1620.
-
-
40 pull requests received 241 new comments.
-
Update f32 tests to handle returning undefined for infinity gpuweb/cts#1655 (30 comments by @ben-clayton, @zoddicus)
-
Add GPURenderPassDescriptor.maxDrawCount test plan gpuweb/cts#1614 (22 comments by @kainino0x, @Kangz, @takahirox)
-
Add WebCodec VideoFrame as input source for GPUExternalTexture cases gpuweb/cts#1711 (19 comments by @kainino0x, @shaoboyan)
-
Succinct the test names of render_pass_descriptor.spec.ts (2/n) gpuweb/cts#1639 (13 comments by @Gyuyoung, @kainino0x)
-
Implement val:render_pass_descriptor:timestamp_writes_query_set_type gpuweb/cts#1653 (12 comments by @Gyuyoung, @kainino0x)
-
Enhance val:render_pass_descriptor:depth_stencil_attachment_depth_clear_value gpuweb/cts#1674 (10 comments by @Gyuyoung, @kainino0x)
-
Add tests for getCurrentTexture and getPreferredCanvasFormat gpuweb/cts#1638 (10 comments by @kainino0x, @toji)
-
Implement val:render_pass_descriptor:occlusion_query_set_type gpuweb/cts#1620 (9 comments by @Gyuyoung, @kainino0x)
-
Describe each test of the validation tests in render|compute pass encoder gpuweb/cts#1573 (9 comments by @kainino0x)
-
Expand readbackFromWebGPUCanvas test with alphaMode gpuweb/cts#1673 (8 comments by @kainino0x, @shrekshao)
-
Add an out-of-range check to the ranges test in writeBuffer.spec.ts gpuweb/cts#1663 (7 comments by @Gyuyoung, @kainino0x)
-
Implement val:render_pass_descriptor:depth_stencil_attachment_depth_clear_value gpuweb/cts#1640 (7 comments by @Gyuyoung, @kainino0x)
-
Test zero levels or layers is invalid in render_pass,resolve:resolve_attachment gpuweb/cts#1600 (7 comments by @Gyuyoung, @kainino0x)
-
Implement api,validation,capability_checks,features,texture_formats:canvas_configuration_view_formats gpuweb/cts#1593 (7 comments by @Gyuyoung, @kainino0x)
-
Guard against writeTimestamp() calls without the proper feature gpuweb/cts#1601 (7 comments by @austinEng, @toji)
-
Add reftests: canvas use as image source with opaque/premultiplied alpha mode gpuweb/cts#1657 (6 comments by @kainino0x, @shrekshao)
-
Implement val:render_pass_descriptor:occlusion_query_set_type gpuweb/cts#1652 (6 comments by @Gyuyoung, @kainino0x)
-
Shorten the vertex buffer OOB validation test name gpuweb/cts#1602 (6 comments by @haoxli, @jiangzhaoming, @Jiawei-Shao)
-
Succinct the test names of render_pass_descriptor.spec.ts (3/3) gpuweb/cts#1651 (4 comments by @Gyuyoung, @kainino0x)
-
Implement val:render_pass_descriptor:timestamp_writes_location gpuweb/cts#1630 (4 comments by @Gyuyoung, @kainino0x)
-
Expand map_detach test for buffer should detach on device.destroy gpuweb/cts#1699 (2 comments by @austinEng, @kainino0x)
-
Make `render_pass_descriptor:resolveTarget,usage` test a valid case together gpuweb/cts#1660 (2 comments by @Gyuyoung, @kainino0x)
-
Optimized several expensive methods in texture tests gpuweb/cts#1583 (2 comments by @kainino0x, @toji)
-
DNS: Prototype for how to implement inherited FP accuracy tests gpuweb/cts#1402 (1 comment by @zoddicus)
-
Correct logInterval & log2Interval handling of out of bounds inputs gpuweb/cts#1629 (1 comment by @zoddicus)
-
gpuweb/gpuweb
-
71 issues opened.
-
wgsl: permit certain resource variables to overlap gpuweb/gpuweb#3266, by @dneto0.
-
Object labels require client-shared state gpuweb/gpuweb#3263, by @kainino0x.
-
Clarification of `@builtin` values in structs not used in an entry-point gpuweb/gpuweb#3261, by @dj2.
-
Accuracy of `mix` does not cover all common implementations of linear blend gpuweb/gpuweb#3260, by @zoddicus.
-
wgsl: Add limits to avoid nuisance floating point literals gpuweb/gpuweb#3255, by @dneto0.
-
Buffer state validation in unmap() is unnecessary gpuweb/gpuweb#3251, by @kainino0x.
-
Multiple components cannot effectively share a GPUDevice because it is stateful gpuweb/gpuweb#3250, by @kainino0x.
-
Clarifications/examples for uniformity section gpuweb/gpuweb#3246, by @dj2.
-
How destroy buffer in this classic use case ? gpuweb/gpuweb#3241, by @ra1jin.
-
Drop support for macOS 10.12? gpuweb/gpuweb#3238, by @kainino0x.
-
wsgl: disallow repeated use of builtin inputs and builtin outputs gpuweb/gpuweb#3236, by @dneto0.
-
wgsl: compute shaders have no user-defined inputs or outputs gpuweb/gpuweb#3227, by @dneto0.
-
Define exclusivity of `Function Tag` and `Parameter Tag` values. gpuweb/gpuweb#3223, by @dj2.
-
publish-TR-webgpu: Attribute “xlink” not allowed on element “svg” at this point gpuweb/gpuweb#3222, by @kainino0x.
-
Capitolization issue in uniformity gpuweb/gpuweb#3221, by @dj2.
-
Define domain for float builtins and describe errors gpuweb/gpuweb#3220, by @dj2.
-
Minor text changes in Memory Model section gpuweb/gpuweb#3219, by @dj2.
-
Parsing of strange float cases gpuweb/gpuweb#3218, by @dj2.
-
Clarification in limits section gpuweb/gpuweb#3217, by @dj2.
-
Duplicate text in enable section gpuweb/gpuweb#3216, by @dj2.
-
Add link to note about number of locations being in WebGPU API. gpuweb/gpuweb#3215, by @dj2.
-
`sign`'s definition is tautological gpuweb/gpuweb#3211, by @zoddicus.
-
Precision requirements for `atan2` appear to be too restrictive for some backends gpuweb/gpuweb#3209, by @zoddicus.
-
Missing definition of `T` in accessor expressions gpuweb/gpuweb#3208, by @dj2.
-
Allow optional `()`s on `enable`? gpuweb/gpuweb#3207, by @dj2.
-
Clarify abstract floats and undefined values gpuweb/gpuweb#3206, by @dj2.
-
Fix title for `entry point IO types` gpuweb/gpuweb#3205, by @dj2.
-
Clarification on `const` in static access gpuweb/gpuweb#3204, by @dj2.
-
Function attribute duplication gpuweb/gpuweb#3203, by @dj2.
-
Entry point declaration example bug gpuweb/gpuweb#3202, by @dj2.
-
Link features, constraints for shader stages gpuweb/gpuweb#3201, by @dj2.
-
Link WebGPU spec for pipelines in shader stages gpuweb/gpuweb#3200, by @dj2.
-
Clarification of dynamic nature of function calls gpuweb/gpuweb#3199, by @dj2.
-
Clarification on control flow of function calls gpuweb/gpuweb#3198, by @dj2.
-
Clarification on parameter matching statement gpuweb/gpuweb#3197, by @dj2.
-
Clarify function parameter initialization gpuweb/gpuweb#3196, by @dj2.
-
User defined function example typo gpuweb/gpuweb#3195, by @dj2.
-
Update links for user-defined functions gpuweb/gpuweb#3194, by @dj2.
-
Clarification in user-defined functions on attributes on parameters gpuweb/gpuweb#3193, by @dj2.
-
Clarification of WGSL builtin function availability gpuweb/gpuweb#3192, by @dj2.
-
Specify the units for the various `cos`, `sin`, etc builtin functions gpuweb/gpuweb#3191, by @dj2.
-
Add text around `discard` in a `for` `update_part` gpuweb/gpuweb#3188, by @dj2.
-
Grammar update in for statement gpuweb/gpuweb#3186, by @dj2.
-
Buffer Offset issue in buffer-texture copies with depth stencil textures gpuweb/gpuweb#3185, by @Jiawei-Shao.
-
Add support for 3D compressed textures. gpuweb/gpuweb#3183, by @mrshannon.
-
Mini provoking vertex investigation. gpuweb/gpuweb#3180, by @Kangz.
-
wgsl: Make definitions for each builtin value, so we can do things like [=builtin/position=] gpuweb/gpuweb#3177, by @dneto0.
-
Position: WebGPU should specify both a bytecode format and a human-authorable format for WGSL gpuweb/gpuweb#3175, by @jemc.
-
Compound assignment under specified. gpuweb/gpuweb#3169, by @dj2.
-
remove keywords that are context-dependent names gpuweb/gpuweb#3166, by @dneto0.
-
getMappedRange is intrusive to the ArrayBuffer spec gpuweb/gpuweb#3155, by @kainino0x.
-
Clarification of loop identifier lifetime gpuweb/gpuweb#3153, by @dj2.
-
Simplify if condition type precondition text gpuweb/gpuweb#3152, by @dj2.
-
Update if control flow after removal of paren requirements gpuweb/gpuweb#3150, by @dj2.
-
wgsl: language evolution: managing rollout of core (essentially sugar) language features gpuweb/gpuweb#3149, by @dneto0.
-
Constant identifier expression table clarifications gpuweb/gpuweb#3147, by @dj2.
-
wgsl: trunc behaviour is not specified for infinity gpuweb/gpuweb#3146, by @dneto0.
-
Allow `f16` for `!=` comparison gpuweb/gpuweb#3137, by @dj2.
-
Clarify uses of `infinity` in the spec gpuweb/gpuweb#3135, by @dj2.
-
`ldexp` needs clarification gpuweb/gpuweb#3134, by @zoddicus.
-
Replace instances of `integral` with `integer scalar` gpuweb/gpuweb#3133, by @dj2.
-
Update arithmetic expressions with results of over/underflow on abstract types gpuweb/gpuweb#3132, by @dj2.
-
Is `@align(1)` valid? gpuweb/gpuweb#3131, by @dj2.
-
Is an entry point required in WGSL? gpuweb/gpuweb#3130, by @dj2.
-
Make definition of `invalid memory reference` more generic gpuweb/gpuweb#3128, by @dj2.
-
Constructor expressions use `T` without defining `T` gpuweb/gpuweb#3127, by @dj2.
-
Cleanup text around inferring store type when discussing function scope let/var gpuweb/gpuweb#3125, by @dj2.
-
-
66 issues closed.
-
Fix title for `entry point IO types` gpuweb/gpuweb#3205.
-
Add text around `discard` in a `for` `update_part` gpuweb/gpuweb#3188.
-
Include arrays in creation-time constant expressions gpuweb/gpuweb#3056.
-
Capitolization issue in uniformity gpuweb/gpuweb#3221.
-
uniformity analysis + basic tail-branching gpuweb/gpuweb#3007.
-
Specify a minimum `SizeOf` for a runtime array gpuweb/gpuweb#2997.
-
Clarifications for IO-shareable types gpuweb/gpuweb#3033.
-
should WGSL allow shadowing of builtins at module scope gpuweb/gpuweb#2941.
-
`sign`'s definition is tautological gpuweb/gpuweb#3211.
-
How destroy buffer in this classic use case ? gpuweb/gpuweb#3241.
-
Cleanup text around inferring store type when discussing function scope let/var gpuweb/gpuweb#3125.
-
Clarification around usage of module constants gpuweb/gpuweb#3085.
-
Clarifications for initializers for creation-time constants gpuweb/gpuweb#3086.
-
Confusing language around `variable` and `value declaration` gpuweb/gpuweb#3083.
-
Module constant section doesn't mention initializer requirement gpuweb/gpuweb#3084.
-
Clarify lifetime of a variable gpuweb/gpuweb#3080.
-
Clarify `var` initializer in non-private/function address spaces. gpuweb/gpuweb#3077.
-
Extend let declaration description to make clear how it is different from const gpuweb/gpuweb#3063.
-
Clarify that var declarations can infer the store type gpuweb/gpuweb#3076.
-
Extend value declaration to other consequences of no memory locations? gpuweb/gpuweb#3062.
-
Breakdown of var section is confusing gpuweb/gpuweb#3049.
-
Clarify specification around inferring module scoped variable types. gpuweb/gpuweb#3048.
-
Texture binding sampleType for "float or depth" gpuweb/gpuweb#3115.
-
Clarify abstract floats and undefined values gpuweb/gpuweb#3206.
-
Require loops to terminate gpuweb/gpuweb#1987.
-
Q: will implementations provide additional builtin functions/structures? gpuweb/gpuweb#3102.
-
Some limits don't need to be configurable gpuweb/gpuweb#1865.
-
Investigation: kill vs demoteToHelper behavior in native APIs gpuweb/gpuweb#921.
-
Update if control flow after removal of paren requirements gpuweb/gpuweb#3150.
-
wgsl: should `length` builtin have AbstractFloat and f16 overloads? gpuweb/gpuweb#3111.
-
Consider putting most of the writeTexture validation on the device timeline. gpuweb/gpuweb#2680.
-
getMappedRange is intrusive to the ArrayBuffer spec gpuweb/gpuweb#3155.
-
Reading from MSAA depth textures gpuweb/gpuweb#1924.
-
Mini provoking vertex investigation. gpuweb/gpuweb#3180.
-
Add rg11b10ufloat as valid render attachment format gpuweb/gpuweb#2648.
-
Compound assignment under specified. gpuweb/gpuweb#3169.
-
Simplify if condition type precondition text gpuweb/gpuweb#3152.
-
Clarify usage of `@group` and `@binding` decorations gpuweb/gpuweb#3082.
-
Should objects have pointers back to their parents? gpuweb/gpuweb#1497.
-
Missing conditional for fixed-footprint fixed-size array types gpuweb/gpuweb#3022.
-
Pointer and Ref types are missing from sidebar gpuweb/gpuweb#3025.
-
Example: Access Modes for Buffers seems out of place gpuweb/gpuweb#3060.
-
Specify that structures compare by name, not by structure. gpuweb/gpuweb#2994.
-
Specify how pointer types compare? gpuweb/gpuweb#3030.
-
Update `texture_sampler_types` grammar rule to `texture_and_sampler_types` gpuweb/gpuweb#3059.
-
Update Limits.md documentation gpuweb/gpuweb#3027.
-
Add `saturate()` or `clamp01()` gpuweb/gpuweb#3012.
-
Example of unary negation overloads is out of date. gpuweb/gpuweb#2938.
-
Update or remove note about function declarations containing other declarations gpuweb/gpuweb#2936.
-
`textureDimension` missing `i32` overload for `level` parameter gpuweb/gpuweb#2908.
-
Boolean Type rules table still references SPIR-V gpuweb/gpuweb#2973.
-
Spec prose seems to allow `staticAssert` before `global_directive` gpuweb/gpuweb#2931.
-
Clarify `Valid Values` language in the attributes table. gpuweb/gpuweb#2916.
-
textureSampleLevel missing `external` in definition of `t` gpuweb/gpuweb#2896.
-
textureSampleCompareLevel missing params and return documentation. gpuweb/gpuweb#2891.
-
Allow `f16` for `!=` comparison gpuweb/gpuweb#3137.
-
wgsl: extractBits out of alphabetical order in spec text gpuweb/gpuweb#2793.
-
Parameterization missing for `arrayLength` gpuweb/gpuweb#2818.
-
Disambiguate "storage" in WGSL spec gpuweb/gpuweb#2163.
-
in certain contexts, an identifier does not resolve to its declared object gpuweb/gpuweb#2486.
-
wgsl: Identifiers never resolve in certain contexts gpuweb/gpuweb#3113.
-
The validating linear texture data algorithm is overly complex gpuweb/gpuweb#2971.
-
-
4 issues reopened.
-
timestamp-query is unimplementable on TBDR architectures gpuweb/gpuweb#2046, by @haoxli.
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094, by @kainino0x.
-
Make definition of `invalid memory reference` more generic gpuweb/gpuweb#3128, by @dj2.
-
Escape Mechanisms in Identifiers gpuweb/gpuweb#2810, by @kdashg.
-
-
86 issues received 183 new comments.
-
Criteria for pushing the WebGPU and WGSL specification to Candidate Recommendation gpuweb/gpuweb#3107 (12 comments by @fserb, @kainino0x, @Kangz, @kdashg, @tidoust)
-
Buffer Offset issue in buffer-texture copies with depth stencil textures gpuweb/gpuweb#3185 (10 comments by @jenatali, @jimblandy, @kainino0x, @Kangz, @magcius)
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766 (10 comments by @greggman, @kainino0x, @Kangz, @kdashg, @litherum)
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094 (6 comments by @kainino0x, @litherum, @magcius)
-
Clarification of `@builtin` values in structs not used in an entry-point gpuweb/gpuweb#3261 (5 comments by @dj2, @dneto0)
-
Can't enable webgpu unsafe flag on Chrome Canary v102.0.4953.0 gpuweb/gpuweb#2674 (4 comments by @devshgraphicsprogramming, @LifeIsStrange, @willzhai)
-
Make lost devices appear to function as much as possible gpuweb/gpuweb#1629 (4 comments by @greggman, @kainino0x)
-
publish-TR-webgpu: Attribute “xlink” not allowed on element “svg” at this point gpuweb/gpuweb#3222 (3 comments by @kainino0x, @tidoust)
-
Position: WebGPU should specify both a bytecode format and a human-authorable format for WGSL gpuweb/gpuweb#3175 (3 comments by @jemc, @Kangz)
-
Add example in Memory Views section showing a ref type. gpuweb/gpuweb#3053 (3 comments by @dj2, @dneto0)
-
should WGSL allow shadowing of builtins at module scope gpuweb/gpuweb#2941 (3 comments by @dneto0, @kdashg, @litherum)
-
Accuracy of `mix` does not cover all common implementations of linear blend gpuweb/gpuweb#3260 (2 comments by @alan-baker, @dneto0)
-
List of issues to discuss with TC39, WASM, JS engine teams gpuweb/gpuweb#747 (2 comments by @kainino0x)
-
Make definition of `invalid memory reference` more generic gpuweb/gpuweb#3128 (2 comments by @dj2, @dneto0)
-
Should objects have pointers back to their parents? gpuweb/gpuweb#1497 (2 comments by @kdashg, @toji)
-
Update comment about reference types not appearing to a note. gpuweb/gpuweb#3052 (2 comments by @dj2, @dneto0)
-
wgsl: Is/should initializer of module-scope variables limited to creation-time expressions? gpuweb/gpuweb#3081 (2 comments by @kdashg)
-
Specify the units for the various `cos`, `sin`, etc builtin functions gpuweb/gpuweb#3191 (1 comment by @dneto0)
-
wgsl: Add limits to avoid nuisance floating point literals gpuweb/gpuweb#3255 (1 comment by @dneto0)
-
invalid expressions in const-expr and override-expr should be error (no later than pipeline creation time) gpuweb/gpuweb#3253 (1 comment by @dneto0)
-
wgsl: compute shaders have no user-defined inputs or outputs gpuweb/gpuweb#3227 (1 comment by @dneto0)
-
Precision requirements for `atan2` appear to be too restrictive for some backends gpuweb/gpuweb#3209 (1 comment by @dneto0)
-
Specification needs to clarify behavior of indexing of abstract composite type with runtime value gpuweb/gpuweb#3210 (1 comment by @ben-clayton)
-
Q: will implementations provide additional builtin functions/structures? gpuweb/gpuweb#3102 (1 comment by @Kangz)
-
Clarify meaning of `context` in integer example in `TYPE INFERENCE FOR LITERALS` gpuweb/gpuweb#2970 (1 comment by @kdashg)
-
Consider putting most of the writeTexture validation on the device timeline. gpuweb/gpuweb#2680 (1 comment by @kainino0x)
-
Update arithmetic expressions with results of over/underflow on abstract types gpuweb/gpuweb#3132 (1 comment by @dneto0)
-
Confusing language around `variable` and `value declaration` gpuweb/gpuweb#3083 (1 comment by @dneto0)
-
Extend let declaration description to make clear how it is different from const gpuweb/gpuweb#3063 (1 comment by @dneto0)
-
Change `creation-time constant` in Value Declarations to `const`? gpuweb/gpuweb#3061 (1 comment by @dneto0)
-
Structure types should specify that the contained vectors and matrices must be concretized components gpuweb/gpuweb#3032 (1 comment by @dneto0)
-
wgsl: Add some clarification in the `Scalar Types` section that an Abstract is not a scalar type gpuweb/gpuweb#2975 (1 comment by @dneto0)
-
wgsl: should `length` builtin have AbstractFloat and f16 overloads? gpuweb/gpuweb#3111 (1 comment by @kdashg)
-
`textureDimension` missing `i32` overload for `level` parameter gpuweb/gpuweb#2908 (1 comment by @dneto0)
-
in certain contexts, an identifier does not resolve to its declared object gpuweb/gpuweb#2486 (1 comment by @dneto0)
-
-
71 pull requests opened.
-
Fix repeated 'keyword' in explanation of extension names gpuweb/gpuweb#3270, by @dneto0.
-
Add @workgroup_size attributes on compute shader examples gpuweb/gpuweb#3269, by @dneto0.
-
Fix typo in example of user-defined function gpuweb/gpuweb#3268, by @dneto0.
-
clarify lifetime of a declaration in a loop body gpuweb/gpuweb#3267, by @dneto0.
-
Comment out railroad diagram to fix TR publishing gpuweb/gpuweb#3265, by @kainino0x.
-
Simplify device-timeline map states gpuweb/gpuweb#3264, by @kainino0x.
-
Move "Inteprolation" after "Input-output Locations" gpuweb/gpuweb#3262, by @dneto0.
-
Add mermaid diagram support at build time gpuweb/gpuweb#3257, by @kainino0x.
-
Write "identifer" instead of "identifier use" gpuweb/gpuweb#3256, by @dneto0.
-
Make function returns reconverge gpuweb/gpuweb#3254, by @alan-baker.
-
Add mermaid diagram support, with example gpuweb/gpuweb#3252, by @kainino0x.
-
Clarify what location applies to gpuweb/gpuweb#3249, by @dneto0.
-
Simplify 'select' parameterization: scalars can be abstract gpuweb/gpuweb#3248, by @dneto0.
-
Remove confusing statement about identifiers declared inside `loop` gpuweb/gpuweb#3247, by @dneto0.
-
wgsl: Fix result type of distance built-in gpuweb/gpuweb#3242, by @jiangzhaoming.
-
There are four kinds of value declarations. gpuweb/gpuweb#3240, by @jimblandy.
-
Adjust argumentdef style to fit better within the width of the spec gpuweb/gpuweb#3239, by @kainino0x.
-
Clean up description of entry point IO gpuweb/gpuweb#3237, by @dneto0.
-
Add abstract types to scalar types gpuweb/gpuweb#3235, by @alan-baker.
-
Merge floating-point, integer, matrix and vector built-in functions gpuweb/gpuweb#3234, by @alan-baker.
-
Allow expressions in attribute parameters gpuweb/gpuweb#3233, by @alan-baker.
-
Rename staticAssert to static_assert gpuweb/gpuweb#3232, by @alan-baker.
-
Separate subsection for each method on GPUQueue gpuweb/gpuweb#3231, by @kainino0x.
-
Allow using depth textures with 'unfilterable-float' bindings gpuweb/gpuweb#3230, by @kainino0x.
-
Work around xlink validation failure gpuweb/gpuweb#3229, by @kainino0x.
-
Set currentTexture to null on presentation gpuweb/gpuweb#3228, by @kainino0x.
-
Rewrite variables and values for clarity gpuweb/gpuweb#3226, by @alan-baker.
-
Clarify abstract numeric expressions can't produce infinity or NaN gpuweb/gpuweb#3225, by @dneto0.
-
Un-nest steps for mapAsync() gpuweb/gpuweb#3224, by @kainino0x.
-
[editorial] Move abstract numeric types under plain types gpuweb/gpuweb#3214, by @alan-baker.
-
Add abstract arrays gpuweb/gpuweb#3213, by @alan-baker.
-
Specify sign(.) more clearly gpuweb/gpuweb#3212, by @dneto0.
-
Move override-expression definition into its own section gpuweb/gpuweb#3190, by @alan-baker.
-
Rename creation-time constant/expresion/function to const- gpuweb/gpuweb#3189, by @alan-baker.
-
Fix length overloads gpuweb/gpuweb#3184, by @alan-baker.
-
Un-nest timeline steps gpuweb/gpuweb#3182, by @kainino0x.
-
Describe writeTexture as copying the entire source to the device timeline gpuweb/gpuweb#3181, by @kainino0x.
-
[editorial] Add definitions for built-in values gpuweb/gpuweb#3179, by @alan-baker.
-
Fix double word "say say" gpuweb/gpuweb#3178, by @dneto0.
-
Describe discard as demoting to a helper invocation gpuweb/gpuweb#3176, by @alan-baker.
-
Fix wording of if behaviour, re. parenthesization gpuweb/gpuweb#3174, by @dneto0.
-
Simplify 'if' description gpuweb/gpuweb#3173, by @dneto0.
-
LHS of compound assignment has read_write access mode gpuweb/gpuweb#3172, by @dneto0.
-
Delete vague statement about @group and @binding gpuweb/gpuweb#3171, by @dneto0.
-
Add rg11b10ufloat-renderable feature gpuweb/gpuweb#3170, by @kainino0x.
-
Note that objects may hold strong references to parents gpuweb/gpuweb#3168, by @toji.
-
Remove types from reserved word list gpuweb/gpuweb#3167, by @dneto0.
-
Rename to texture_and_sampler_types gpuweb/gpuweb#3165, by @dneto0.
-
Fix note about builtins with texture type as a parameter gpuweb/gpuweb#3164, by @dneto0.
-
Move buffer access mode example gpuweb/gpuweb#3163, by @dneto0.
-
Structure members of vector and matrix type must be concrete gpuweb/gpuweb#3162, by @dneto0.
-
Make 'reference' and 'pointer type' appear in table-of-contents gpuweb/gpuweb#3161, by @dneto0.
-
Make mapping ArrayBuffers non-detachable gpuweb/gpuweb#3160, by @kainino0x.
-
Specify when pointer types are the same gpuweb/gpuweb#3159, by @dneto0.
-
Fixed-footprint types: add remark about fixed-sized arrays gpuweb/gpuweb#3158, by @dneto0.
-
@size attrib only applied to member types with creation-fixed footprint gpuweb/gpuweb#3157, by @dneto0.
-
structure types compare by name, not structure gpuweb/gpuweb#3156, by @dneto0.
-
Allow shadowing of predeclared objects gpuweb/gpuweb#3154, by @dneto0.
-
Document the dependency on py-tree-sitter gpuweb/gpuweb#3148, by @dneto0.
-
Clarify: function is only kind of decl that can contain other decls gpuweb/gpuweb#3145, by @dneto0.
-
Clarify directives come first gpuweb/gpuweb#3144, by @dneto0.
-
Clarify attribute integer paramters validity gpuweb/gpuweb#3143, by @dneto0.
-
Remove references to OpConstantTrue, OpConstantFalse gpuweb/gpuweb#3142, by @dneto0.
-
Allow level to be i32 in textureDimensions gpuweb/gpuweb#3141, by @dneto0.
-
Add external texture to description of 't' for textureSampleLevel gpuweb/gpuweb#3140, by @dneto0.
-
Add f16 parameterization for != gpuweb/gpuweb#3139, by @dneto0.
-
Describe params and return for textureSampleCompareLevel gpuweb/gpuweb#3138, by @dneto0.
-
Move extractBits into alphabetical order gpuweb/gpuweb#3136, by @dneto0.
-
Generalize invalid memory reference to any originating variable gpuweb/gpuweb#3129, by @dneto0.
-
Say that shaders should terminate, or UA may lose the device gpuweb/gpuweb#3126, by @dneto0.
-
Describe "mapped range ArrayBuffers" in terms of copies gpuweb/gpuweb#3124, by @kainino0x.
-
-
67 pull requests merged.
-
Comment out railroad diagram to fix TR publishing gpuweb/gpuweb#3265.
-
Add mermaid diagram support at build time gpuweb/gpuweb#3257.
-
Move "Inteprolation" after "Input-output Locations" gpuweb/gpuweb#3262.
-
Write "identifer" instead of "identifier use" gpuweb/gpuweb#3256.
-
Make function returns reconverge gpuweb/gpuweb#3254.
-
@size attrib only applied to member types with creation-fixed footprint gpuweb/gpuweb#3157.
-
Simplify 'select' parameterization: scalars can be abstract gpuweb/gpuweb#3248.
-
Clarify what location applies to gpuweb/gpuweb#3249.
-
Clean up description of entry point IO gpuweb/gpuweb#3237.
-
Allow expressions in attribute parameters gpuweb/gpuweb#3233.
-
Add abstract types to scalar types gpuweb/gpuweb#3235.
-
Remove types from reserved word list gpuweb/gpuweb#3167.
-
Rename staticAssert to static_assert gpuweb/gpuweb#3232.
-
Specify sign(.) more clearly gpuweb/gpuweb#3212.
-
There are four kinds of value declarations. gpuweb/gpuweb#3240.
-
wgsl: Fix result type of distance built-in gpuweb/gpuweb#3242.
-
Rewrite variables and values for clarity gpuweb/gpuweb#3226.
-
Merge floating-point, integer, matrix and vector built-in functions gpuweb/gpuweb#3234.
-
Set currentTexture to null on presentation gpuweb/gpuweb#3228.
-
Un-nest steps for mapAsync() gpuweb/gpuweb#3224.
-
Work around xlink validation failure gpuweb/gpuweb#3229.
-
Clarify abstract numeric expressions can't produce infinity or NaN gpuweb/gpuweb#3225.
-
[editorial] Move abstract numeric types under plain types gpuweb/gpuweb#3214.
-
Un-nest timeline steps gpuweb/gpuweb#3182.
-
Describe writeTexture as copying the entire source to the device timeline gpuweb/gpuweb#3181.
-
Add abstract arrays gpuweb/gpuweb#3213.
-
Move override-expression definition into its own section gpuweb/gpuweb#3190.
-
Say that shaders should terminate, or UA may lose the device gpuweb/gpuweb#3126.
-
Rename creation-time constant/expresion/function to const- gpuweb/gpuweb#3189.
-
Describe discard as demoting to a helper invocation gpuweb/gpuweb#3176.
-
Fix wording of if behaviour, re. parenthesization gpuweb/gpuweb#3174.
-
Fix length overloads gpuweb/gpuweb#3184.
-
[editorial] Add definitions for built-in values gpuweb/gpuweb#3179.
-
Allow inter-stage variable subsetting gpuweb/gpuweb#3122.
-
Make mapping ArrayBuffers non-detachable gpuweb/gpuweb#3160.
-
Allow BGL entries for multisampled:true "depth" textures gpuweb/gpuweb#3072.
-
Fix double word "say say" gpuweb/gpuweb#3178.
-
Add rg11b10ufloat-renderable feature gpuweb/gpuweb#3170.
-
LHS of compound assignment has read_write access mode gpuweb/gpuweb#3172.
-
Simplify 'if' description gpuweb/gpuweb#3173.
-
Delete vague statement about @group and @binding gpuweb/gpuweb#3171.
-
Note that objects may hold strong references to parents gpuweb/gpuweb#3168.
-
Structure members of vector and matrix type must be concrete gpuweb/gpuweb#3162.
-
Fixed-footprint types: add remark about fixed-sized arrays gpuweb/gpuweb#3158.
-
Make 'reference' and 'pointer type' appear in table-of-contents gpuweb/gpuweb#3161.
-
Move buffer access mode example gpuweb/gpuweb#3163.
-
Fix note about builtins with texture type as a parameter gpuweb/gpuweb#3164.
-
structure types compare by name, not structure gpuweb/gpuweb#3156.
-
Specify when pointer types are the same gpuweb/gpuweb#3159.
-
Rename to texture_and_sampler_types gpuweb/gpuweb#3165.
-
Document the dependency on py-tree-sitter gpuweb/gpuweb#3148.
-
Allow depth textures with cube/array/multisampled types gpuweb/gpuweb#3070.
-
Correspondence Reference: Limits gpuweb/gpuweb#3088.
-
Allow shadowing of predeclared objects gpuweb/gpuweb#3154.
-
Add saturate floating point builtin function gpuweb/gpuweb#3068.
-
Clarify: function is only kind of decl that can contain other decls gpuweb/gpuweb#3145.
-
Allow level to be i32 in textureDimensions gpuweb/gpuweb#3141.
-
Remove references to OpConstantTrue, OpConstantFalse gpuweb/gpuweb#3142.
-
Clarify directives come first gpuweb/gpuweb#3144.
-
Clarify attribute integer paramters validity gpuweb/gpuweb#3143.
-
Add external texture to description of 't' for textureSampleLevel gpuweb/gpuweb#3140.
-
Describe params and return for textureSampleCompareLevel gpuweb/gpuweb#3138.
-
Add f16 parameterization for != gpuweb/gpuweb#3139.
-
Move extractBits into alphabetical order gpuweb/gpuweb#3136.
-
Generalize invalid memory reference to any originating variable gpuweb/gpuweb#3129.
-
In certain places, identifiers don't resolve to declarations gpuweb/gpuweb#3117.
-
Correspondence References: Resources gpuweb/gpuweb#3093.
-
-
4 pull requests closed.
-
Add mermaid diagram support, with example gpuweb/gpuweb#3252.
-
Remove confusing statement about identifiers declared inside `loop` gpuweb/gpuweb#3247.
-
Add pointers back to parent objects for most interfaces gpuweb/gpuweb#2484.
-
Describe "mapped range ArrayBuffers" in terms of copies gpuweb/gpuweb#3124.
-
-
1 pull request reopened.
-
Structure members of vector and matrix type must be concrete gpuweb/gpuweb#3162, by @dneto0.
-
-
39 pull requests received 190 new comments.
-
Rewrite variables and values for clarity gpuweb/gpuweb#3226 (26 comments by @alan-baker, @dj2, @dneto0)
-
Describe discard as demoting to a helper invocation gpuweb/gpuweb#3176 (18 comments by @alan-baker, @dj2, @dneto0)
-
Move override-expression definition into its own section gpuweb/gpuweb#3190 (16 comments by @alan-baker, @dneto0)
-
Add mermaid diagram support at build time gpuweb/gpuweb#3257 (14 comments by @dj2, @kainino0x, @tidoust)
-
Add mermaid diagram support, with example gpuweb/gpuweb#3252 (8 comments by @dneto0, @kainino0x, @mehmetoguzderin, @tidoust)
-
Say that shaders should terminate, or UA may lose the device gpuweb/gpuweb#3126 (7 comments by @dneto0, @kainino0x, @litherum, @toji)
-
Describe "mapped range ArrayBuffers" in terms of copies gpuweb/gpuweb#3124 (7 comments by @jimblandy, @kainino0x)
-
Rename creation-time constant/expresion/function to const- gpuweb/gpuweb#3189 (2 comments by @dneto0)
-
wgsl: fold expression semantics text into definition of type rule gpuweb/gpuweb#2444 (1 comment by @dneto0)
-
@size attrib only applied to member types with creation-fixed footprint gpuweb/gpuweb#3157 (1 comment by @dneto0)
-
Remove confusing statement about identifiers declared inside `loop` gpuweb/gpuweb#3247 (1 comment by @dneto0)
-
Merge floating-point, integer, matrix and vector built-in functions gpuweb/gpuweb#3234 (1 comment by @jiangzhaoming)
-
Note that objects may hold strong references to parents gpuweb/gpuweb#3168 (1 comment by @kainino0x)
-
Fixed-footprint types: add remark about fixed-sized arrays gpuweb/gpuweb#3158 (1 comment by @dneto0)
-
Clarify: function is only kind of decl that can contain other decls gpuweb/gpuweb#3145 (1 comment by @dneto0)
-
Generalize invalid memory reference to any originating variable gpuweb/gpuweb#3129 (1 comment by @alan-baker)
-
TR publications
-
37 documents published.
-
WebGPU published 19 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 18 times as a Working Draft. History of WebGPU Shading Language
-
Participants
-
1 participant left.
-
Robin Morisset (Apple Inc.) left.
-
GitHub
gpuweb/cts
-
33 issues opened.
-
wgsl: length(AbstractFloat) yields f32, not a vector type gpuweb/cts#1588, by @dneto0.
-
wgsl: comparison operators overloads with AbstractInt, AbstractFloat yield bool gpuweb/cts#1587, by @dneto0.
-
val: createTexture: usage 0 is invalid gpuweb/cts#1576, by @Gyuyoung.
-
wgsl: Validation module scoped variables gpuweb/cts#1570, by @dj2.
-
wgsl: Execution tests for var initialization gpuweb/cts#1565, by @dj2.
-
val: createPipelineLayout gpuweb/cts#1563, by @Gyuyoung.
-
Make timeout deadlines configurable gpuweb/cts#1562, by @austinEng.
-
canvas read as an image source alpha mode behavior update gpuweb/cts#1561, by @shrekshao.
-
wgsl: validate creation-time constants gpuweb/cts#1559, by @dj2.
-
wgsl: execution test for override declaration gpuweb/cts#1558, by @dj2.
-
wgsl: Validate `override` declarations gpuweb/cts#1557, by @dj2.
-
val: createComputePipeline gpuweb/cts#1555, by @shrekshao.
-
val: render_pipeline/inter_stage gpuweb/cts#1554, by @shrekshao.
-
val: createRenderPipeline gpuweb/cts#1553, by @shrekshao.
-
wgsl: Validate let declarations gpuweb/cts#1551, by @dj2.
-
wgsl: validate parsing of type aliases gpuweb/cts#1548, by @dj2.
-
wgsl: validate parsing or storage texture types gpuweb/cts#1546, by @dj2.
-
wgsl: validate parsing of sampled texture types gpuweb/cts#1545, by @dj2.
-
wgsl: Validate texture data conversions gpuweb/cts#1544, by @dj2.
-
wgsl: validate extra components when reading / writing texture data gpuweb/cts#1543, by @dj2.
-
wgsl: using a non-filterable texture with a filtering sampler gpuweb/cts#1542, by @dj2.
-
wgsl: Pointer execution tests gpuweb/cts#1540, by @dj2.
-
wgsl: Assignment with references and pointers gpuweb/cts#1539, by @dj2.
-
wgsl: Access mode restrictions gpuweb/cts#1538, by @dj2.
-
Add operation test about clearStencilValue gpuweb/cts#1524, by @Jiawei-Shao.
-
Add tests about creating multisampled textures with RENDER_ATTACHMENT usage gpuweb/cts#1521, by @Jiawei-Shao.
-
Ensure shader creation-time evaluation gpuweb/cts#1519, by @ben-clayton.
-
Add abstract-numeric `ScalarType`s gpuweb/cts#1517, by @ben-clayton.
-
Refactor `oneULP` utility into `oneULP` + `oneULPImpl` gpuweb/cts#1516, by @zoddicus.
-
Add missing RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/cts#1508, by @Jiawei-Shao.
-
wgsl: validate struct parsing gpuweb/cts#1501, by @dj2.
-
12 issues closed.
-
val: createTexture: usage 0 is invalid gpuweb/cts#1576.
-
op: buffers,map_ArrayBuffer gpuweb/cts#1033.
-
Make timeout deadlines configurable gpuweb/cts#1562.
-
Refactor `oneULP` utility into `oneULP` + `oneULPImpl` gpuweb/cts#1516.
-
Add operation test about clearStencilValue gpuweb/cts#1524.
-
val: Resource usage validation for textures/buffers gpuweb/cts#905.
-
wgsl: execution test for shift* methods gpuweb/cts#1244.
-
val: cmds,render,draw vertex access tests gpuweb/cts#906.
-
-
1 issue reopened.
-
19 issues received 37 new comments.
-
val: capability_checks,features errors should be exceptions gpuweb/cts#919 (5 comments by @jchen10, @kainino0x)
-
val: capability_checks,features,texture_formats for GPUTextureDescriptor.viewFormats gpuweb/cts#920 (4 comments by @Gyuyoung, @kainino0x)
-
val: capability_checks,features,texture_formats for GPUTextureViewDescriptor and GPUCanvasConfiguration gpuweb/cts#902 (3 comments by @Gyuyoung, @kainino0x)
-
Add missing RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/cts#1508 (3 comments by @jiangzhaoming, @Jiawei-Shao, @kainino0x)
-
val: cmds,buffer_texture_copies should move to image_copy gpuweb/cts#1009 (2 comments by @Gyuyoung, @kainino0x)
-
val: Update validation tests to enforce per-stage and per-pipeline-layout limits on BGL creation gpuweb/cts#230 (1 comment by @kainino0x)
-
Add tests about creating multisampled textures with RENDER_ATTACHMENT usage gpuweb/cts#1521 (1 comment by @kainino0x)
-
-
65 pull requests opened.
-
Convert `inverseSqrt` tests to use interval framework gpuweb/cts#1597, by @zoddicus.
-
Convert `min` tests to use interval framework gpuweb/cts#1596, by @zoddicus.
-
Convert `max` tests to use interval framework gpuweb/cts#1595, by @zoddicus.
-
Convert subtraction tests to use interval framework gpuweb/cts#1594, by @zoddicus.
-
Convert multiplication tests to use interval framework gpuweb/cts#1591, by @zoddicus.
-
Convert division tests to use interval framework gpuweb/cts#1590, by @zoddicus.
-
Convert `atan2` tests to use interval framework gpuweb/cts#1586, by @zoddicus.
-
Remove storage write tests gpuweb/cts#1585, by @dj2.
-
Implement the writeTimestamp test to expect an exception gpuweb/cts#1584, by @Gyuyoung.
-
Optimized several expensive methods in texture tests gpuweb/cts#1583, by @toji.
-
Convert addition tests to use interval framework gpuweb/cts#1582, by @zoddicus.
-
Convert `sin` tests to use interval framework gpuweb/cts#1581, by @zoddicus.
-
Convert negation tests to new interval framework gpuweb/cts#1580, by @zoddicus.
-
Convert `log2` tests to using new interval framework gpuweb/cts#1579, by @zoddicus.
-
Convert `log` tests to using new interval framework gpuweb/cts#1578, by @zoddicus.
-
Replace `;` with `,` in WGSL structs. gpuweb/cts#1577, by @dj2.
-
Update 'zero_size' test to check 'usage' field gpuweb/cts#1575, by @Gyuyoung.
-
Describe each test of the validation tests in render|compute pass encoder gpuweb/cts#1573, by @Gyuyoung.
-
Move buffer_texture_copies.spec.ts from encoding/cmds to image_copy path gpuweb/cts#1572, by @Gyuyoung.
-
Update the createQuerySet test of query_types.spec.ts to expect an exception gpuweb/cts#1571, by @Gyuyoung.
-
Fix val:render_pipeline,multisample_state,alpha_to_coverage,sample_mask gpuweb/cts#1569, by @shrekshao.
-
Partially revert #1529 gpuweb/cts#1568, by @toji.
-
Partially revert #1532 gpuweb/cts#1567, by @toji.
-
Fix texture_descriptor_view_formats capability check tests gpuweb/cts#1566, by @toji.
-
Describe the number_of_bind_group_layouts_exceeds_the_maximum_value test gpuweb/cts#1564, by @Gyuyoung.
-
Describe the number_of_dynamic_buffers_exceeds_the_maximum_value test gpuweb/cts#1560, by @Gyuyoung.
-
createComputePipeline review, regroup gpuweb/cts#1552, by @shrekshao.
-
Add escape hatch for precisely expressing expectations gpuweb/cts#1550, by @zoddicus.
-
validation/render_pipeline/inter_stage.spec.ts gpuweb/cts#1549, by @shrekshao.
-
Add tests for container type reflection attributes gpuweb/cts#1547, by @toji.
-
Add basic requestAdapterInfo test gpuweb/cts#1541, by @toji.
-
Remove redundant entry point tests gpuweb/cts#1536, by @shrekshao.
-
Convert `floor` tests to use new interval framework gpuweb/cts#1535, by @zoddicus.
-
Convert `exp2` tests to use new interval framework gpuweb/cts#1534, by @zoddicus.
-
Convert `exp` tests to use new interval framework gpuweb/cts#1533, by @zoddicus.
-
Update the texture format tests of GPUColorTargetState to expect exceptions gpuweb/cts#1531, by @Gyuyoung.
-
Convert `cos` tests to using new interval framework gpuweb/cts#1530, by @zoddicus.
-
Update storage texture format tests to expect exceptions gpuweb/cts#1529, by @Gyuyoung.
-
Use two spaces indentation in struct gpuweb/cts#1528, by @Gyuyoung.
-
Do not use `write` access mode for storage buffers gpuweb/cts#1527, by @jrprice.
-
Use commas for struct member delimeters gpuweb/cts#1526, by @jrprice.
-
Add operation tests about stencilClearValue gpuweb/cts#1525, by @Jiawei-Shao.
-
Test creating multisampled textures with RENDER_ATTACHMENT usage gpuweb/cts#1522, by @Jiawei-Shao.
-
Review createRenderPipeline.spec.ts gpuweb/cts#1520, by @shrekshao.
-
Add wrapper around ULP calculation to handle flush vs no flush gpuweb/cts#1518, by @zoddicus.
-
Const tests gpuweb/cts#1515, by @ben-clayton.
-
[wgsl] use subcases in texture tests gpuweb/cts#1514, by @dj2.
-
Remove references to depth24unorm-stencil8 gpuweb/cts#1513, by @toji.
-
Update @stage attributes. gpuweb/cts#1512, by @dj2.
-
Add api,validation,resource_usages,buffer,in_pass_misc:* - Part II gpuweb/cts#1511, by @Jiawei-Shao.
-
Change Scalar.toString() to include hex bits gpuweb/cts#1510, by @zoddicus.
-
Add api,validation,resource_usages,buffer,in_pass_misc:* - Part I gpuweb/cts#1509, by @Jiawei-Shao.
-
Specify RENDER_ATTACHMENT usage when creating multisampled texture gpuweb/cts#1507, by @Jiawei-Shao.
-
Roll @webgpu/types to 0.1.18 gpuweb/cts#1506, by @kainino0x.
-
Convert `ceil` tests to using new interval framework gpuweb/cts#1505, by @zoddicus.
-
Convert `atan` tests to use interval framework gpuweb/cts#1504, by @zoddicus.
-
Convert `abs` tests to use interval framework gpuweb/cts#1503, by @zoddicus.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part V gpuweb/cts#1502, by @Jiawei-Shao.
-
Implement new F32Interval class for floating point tests gpuweb/cts#1500, by @zoddicus.
-
Update texture format tests to expect exceptions gpuweb/cts#1499, by @toji.
-
-
62 pull requests merged.
-
Convert `min` tests to use interval framework gpuweb/cts#1596.
-
Convert multiplication tests to use interval framework gpuweb/cts#1591.
-
Convert division tests to use interval framework gpuweb/cts#1590.
-
Convert `max` tests to use interval framework gpuweb/cts#1595.
-
Convert `atan2` tests to use interval framework gpuweb/cts#1586.
-
Convert addition tests to use interval framework gpuweb/cts#1582.
-
Implement the writeTimestamp test to expect an exception gpuweb/cts#1584.
-
Convert `log2` tests to using new interval framework gpuweb/cts#1579.
-
Convert `log` tests to using new interval framework gpuweb/cts#1578.
-
Remove storage write tests gpuweb/cts#1585.
-
Convert `sin` tests to use interval framework gpuweb/cts#1581.
-
Convert negation tests to new interval framework gpuweb/cts#1580.
-
Update 'zero_size' test to check 'usage' field gpuweb/cts#1575.
-
Convert `floor` tests to use new interval framework gpuweb/cts#1535.
-
Convert `exp2` tests to use new interval framework gpuweb/cts#1534.
-
Convert `exp` tests to use new interval framework gpuweb/cts#1533.
-
Replace `;` with `,` in WGSL structs. gpuweb/cts#1577.
-
Add tests for container type reflection attributes gpuweb/cts#1547.
-
Fix texture_descriptor_view_formats capability check tests gpuweb/cts#1566.
-
Partially revert #1532 gpuweb/cts#1567.
-
Partially revert #1529 gpuweb/cts#1568.
-
createComputePipeline review, regroup gpuweb/cts#1552.
-
validation/render_pipeline/inter_stage.spec.ts gpuweb/cts#1549.
-
Add basic requestAdapterInfo test gpuweb/cts#1541.
-
Remove redundant entry point tests gpuweb/cts#1536.
-
Convert `cos` tests to using new interval framework gpuweb/cts#1530.
-
Convert `ceil` tests to using new interval framework gpuweb/cts#1505.
-
Review createRenderPipeline.spec.ts gpuweb/cts#1520.
-
Update storage texture format tests to expect exceptions gpuweb/cts#1529.
-
Use two spaces indentation in struct gpuweb/cts#1528.
-
Use commas for struct member delimeters gpuweb/cts#1526.
-
Do not use `write` access mode for storage buffers gpuweb/cts#1527.
-
Convert `atan` tests to use interval framework gpuweb/cts#1504.
-
Add wrapper around ULP calculation to handle flush vs no flush gpuweb/cts#1518.
-
Add operation tests about stencilClearValue gpuweb/cts#1525.
-
Test creating multisampled textures with RENDER_ATTACHMENT usage gpuweb/cts#1522.
-
Convert `abs` tests to use interval framework gpuweb/cts#1503.
-
Const tests gpuweb/cts#1515.
-
Change Scalar.toString() to include hex bits gpuweb/cts#1510.
-
[wgsl] use subcases in texture tests gpuweb/cts#1514.
-
Add api,validation,resource_usages,buffer,in_pass_misc:* - Part II gpuweb/cts#1511.
-
Update @stage attributes. gpuweb/cts#1512.
-
Remove references to depth24unorm-stencil8 gpuweb/cts#1513.
-
Implement new F32Interval class for floating point tests gpuweb/cts#1500.
-
Add api,validation,resource_usages,buffer,in_pass_misc:* - Part I gpuweb/cts#1509.
-
Specify RENDER_ATTACHMENT usage when creating multisampled texture gpuweb/cts#1507.
-
Roll @webgpu/types to 0.1.18 gpuweb/cts#1506.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part V gpuweb/cts#1502.
-
Update texture format tests to expect exceptions gpuweb/cts#1499.
-
Factor copyToTexture:canvas tests common code. gpuweb/cts#1497.
-
-
1 pull request closed.
-
Add escape hatch for precisely expressing expectations gpuweb/cts#1550.
-
-
41 pull requests received 204 new comments.
-
Implement new F32Interval class for floating point tests gpuweb/cts#1500 (38 comments by @ben-clayton, @zoddicus)
-
validation/render_pipeline/inter_stage.spec.ts gpuweb/cts#1549 (15 comments by @kainino0x, @shrekshao)
-
Change Scalar.toString() to include hex bits gpuweb/cts#1510 (13 comments by @amaiorano, @ben-clayton, @zoddicus)
-
Implement api,validation,capability_checks,features,texture_formats:canvas_configuration gpuweb/cts#1537 (11 comments by @Gyuyoung, @kainino0x)
-
Convert `abs` tests to use interval framework gpuweb/cts#1503 (11 comments by @ben-clayton, @Leandromanja, @zoddicus)
-
Describe each test of the validation tests in render|compute pass encoder gpuweb/cts#1573 (10 comments by @Gyuyoung, @kainino0x)
-
Convert multiplication tests to use interval framework gpuweb/cts#1591 (8 comments by @dneto0, @zoddicus)
-
Convert `atan` tests to use interval framework gpuweb/cts#1504 (7 comments by @ben-clayton, @zoddicus)
-
Optimized several expensive methods in texture tests gpuweb/cts#1583 (5 comments by @austinEng, @kainino0x, @toji)
-
Convert `exp` tests to use new interval framework gpuweb/cts#1533 (5 comments by @ben-clayton, @zoddicus)
-
Implement api,validation,capability_checks,features,texture_formats:texture_view_descriptor (#902) gpuweb/cts#1523 (5 comments by @Gyuyoung, @kainino0x)
-
Update the createQuerySet test of query_types.spec.ts to expect an exception gpuweb/cts#1571 (4 comments by @Gyuyoung, @kainino0x, @toji)
-
Add wrapper around ULP calculation to handle flush vs no flush gpuweb/cts#1518 (4 comments by @ben-clayton, @zoddicus)
-
Add operation tests about stencilClearValue gpuweb/cts#1525 (4 comments by @jiangzhaoming, @Jiawei-Shao)
-
Change importExternalTexture,destroy to importExternalTexture,expired gpuweb/cts#1474 (4 comments by @Jiawei-Shao, @shaoboyan)
-
Add tests for container type reflection attributes gpuweb/cts#1547 (3 comments by @kainino0x, @toji)
-
Move buffer_texture_copies.spec.ts from encoding/cmds to image_copy path gpuweb/cts#1572 (2 comments by @Gyuyoung, @kainino0x)
-
Convert `cos` tests to using new interval framework gpuweb/cts#1530 (2 comments by @ben-clayton, @zoddicus)
-
Describe the number_of_dynamic_buffers_exceeds_the_maximum_value test gpuweb/cts#1560 (1 comment by @Gyuyoung)
-
Implement api,validation,capability_checks,features,texture_formats:texture_descriptor_view_formats gpuweb/cts#1556 (1 comment by @Gyuyoung)
-
Update the color and depth stencil formats of RenderBundleEncoder tests to expect exceptions gpuweb/cts#1532 (1 comment by @toji)
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part IV gpuweb/cts#1496 (1 comment by @Jiawei-Shao)
-
gpuweb/gpuweb
-
70 issues opened.
-
Texture binding sampleType for "float or depth" gpuweb/gpuweb#3115, by @kainino0x.
-
wgsl: Identifiers never resolve in certain contexts gpuweb/gpuweb#3113, by @dneto0.
-
wgsl: should `length` builtin have AbstractFloat and f16 overloads? gpuweb/gpuweb#3111, by @dneto0.
-
Criteria for pushing the WebGPU and WGSL specification to Candidate Recommendation gpuweb/gpuweb#3107, by @Kangz.
-
wgsl: log accuracy is in terms of an interval of x, not an interval of log(x) gpuweb/gpuweb#3104, by @dneto0.
-
Q: will implementations provide additional builtin functions/structures? gpuweb/gpuweb#3102, by @pes10k.
-
Fingerprinting Surface gpuweb/gpuweb#3101, by @pes10k.
-
.github/workflows/versions.yml gpuweb/gpuweb#3095, by @callree.
-
Clarifications for initializers for creation-time constants gpuweb/gpuweb#3086, by @dj2.
-
Clarification around usage of module constants gpuweb/gpuweb#3085, by @dj2.
-
Module constant section doesn't mention initializer requirement gpuweb/gpuweb#3084, by @dj2.
-
Confusing language around `variable` and `value declaration` gpuweb/gpuweb#3083, by @dj2.
-
Clarify usage of `@group` and `@binding` decorations gpuweb/gpuweb#3082, by @dj2.
-
Clarify lifetime of a variable gpuweb/gpuweb#3080, by @dj2.
-
Clarify `var` initializer in non-private/function address spaces. gpuweb/gpuweb#3077, by @dj2.
-
Clarify that var declarations can infer the store type gpuweb/gpuweb#3076, by @dj2.
-
Move comment about named values and abstracts. gpuweb/gpuweb#3075, by @dj2.
-
Clarification around `vector` and `matrix` for creation-time constant gpuweb/gpuweb#3074, by @dj2.
-
Value of reflection attributes if validation fails gpuweb/gpuweb#3064, by @toji.
-
Extend let declaration description to make clear how it is different from const gpuweb/gpuweb#3063, by @dj2.
-
Extend value declaration to other consequences of no memory locations? gpuweb/gpuweb#3062, by @dj2.
-
Change `creation-time constant` in Value Declarations to `const`? gpuweb/gpuweb#3061, by @dj2.
-
Example: Access Modes for Buffers seems out of place gpuweb/gpuweb#3060, by @dj2.
-
Update `texture_sampler_types` grammar rule to `texture_and_sampler_types` gpuweb/gpuweb#3059, by @dj2.
-
Update language about the texture being the first parameter to the texture builtins gpuweb/gpuweb#3058, by @dj2.
-
Dynamic error description for read-modify-write atomics could use more explanation. gpuweb/gpuweb#3057, by @dj2.
-
Include arrays in creation-time constant expressions gpuweb/gpuweb#3056, by @ben-clayton.
-
Add example in Memory Views section showing a ref type. gpuweb/gpuweb#3053, by @dj2.
-
Update comment about reference types not appearing to a note. gpuweb/gpuweb#3052, by @dj2.
-
Structure member layout offset placement doesn't not mention padding. gpuweb/gpuweb#3051, by @dj2.
-
Structure member layout either list the align restrictions or remove the size restrictions gpuweb/gpuweb#3050, by @dj2.
-
Breakdown of var section is confusing gpuweb/gpuweb#3049, by @dj2.
-
Clarify specification around inferring module scoped variable types. gpuweb/gpuweb#3048, by @dj2.
-
Clarify note about write-only storage textures in 4.4.6 Address Spaces. gpuweb/gpuweb#3047, by @dj2.
-
Add a section for `constructable host-shareable types` ? gpuweb/gpuweb#3046, by @dj2.
-
Add text and example around IO locations spanning multiple locations gpuweb/gpuweb#3045, by @dj2.
-
builtins: reorder to put Overload, Parameterization before Description and notes gpuweb/gpuweb#3044, by @dneto0.
-
Questions about index continuity where setBindGroup(index, ...) gpuweb/gpuweb#3042, by @ghost.
-
Separate read-only state for stencil and depth aspects. gpuweb/gpuweb#3036, by @brendan-duncan.
-
Clarification around `host-shareable types` gpuweb/gpuweb#3035, by @dj2.
-
Overloading of the word `pipeline` gpuweb/gpuweb#3034, by @dj2.
-
Clarifications for IO-shareable types gpuweb/gpuweb#3033, by @dj2.
-
Clarify storable type concrete components for vectors and matrices. gpuweb/gpuweb#3031, by @dj2.
-
Specify how pointer types compare? gpuweb/gpuweb#3030, by @dj2.
-
Clarifications around memory locations gpuweb/gpuweb#3029, by @dj2.
-
Update Limits.md documentation gpuweb/gpuweb#3027, by @kainino0x.
-
Clarify use of variable in memory overlap and pointers sections gpuweb/gpuweb#3026, by @dj2.
-
Pointer and Ref types are missing from sidebar gpuweb/gpuweb#3025, by @dj2.
-
Clarification of note in fixed-footprint types section gpuweb/gpuweb#3024, by @dj2.
-
Clarify note about array type with overrideable constant sizing gpuweb/gpuweb#3023, by @dj2.
-
Missing conditional for fixed-footprint fixed-size array types gpuweb/gpuweb#3022, by @dj2.
-
Provide stronger control over physical alignments of buffers and bindings gpuweb/gpuweb#3021, by @kainino0x.
-
Revive depth24unorm(-stencil8) gpuweb/gpuweb#3017, by @kainino0x.
-
Mesh shader support gpuweb/gpuweb#3015, by @rishflab.
-
Try-catching getMappedRange is useful but shouldn't be gpuweb/gpuweb#3013, by @kainino0x.
-
Add `saturate()` or `clamp01()` gpuweb/gpuweb#3012, by @jespertheend.
-
uniformity analysis + basic tail-branching gpuweb/gpuweb#3007, by @magcius.
-
Is an atomic a composite type? gpuweb/gpuweb#3004, by @dj2.
-
Is there a method similar to "gl.getActiveXXX" in ShaderModule? gpuweb/gpuweb#3003, by @axmand.
-
16-bit normalized texture support gpuweb/gpuweb#3001, by @magcius.
-
wgsl: document how much of an alignment request is checked by the API gpuweb/gpuweb#2999, by @dneto0.
-
Update list of attributes for structure members gpuweb/gpuweb#2998, by @dj2.
-
Specify a minimum `SizeOf` for a runtime array gpuweb/gpuweb#2997, by @dj2.
-
Update `Pipeline input and output interface` section to reference `@builtin` and `@location` gpuweb/gpuweb#2996, by @dj2.
-
Specify that structures compare by name, not by structure. gpuweb/gpuweb#2994, by @dj2.
-
-
30 issues closed.
-
Clarify the states and lifetime of GPUCommandBuffer gpuweb/gpuweb#1294.
-
wgsl: log accuracy is in terms of an interval of x, not an interval of log(x) gpuweb/gpuweb#3104.
-
Fake transfers for getMappedRange-created ArrayBuffers? gpuweb/gpuweb#2072.
-
[TAG review] Have a note suggesting surface a warning about power usage? gpuweb/gpuweb#2947.
-
.github/workflows/versions.yml gpuweb/gpuweb#3095.
-
Consider formatting algorithm calls like function calls() gpuweb/gpuweb#768.
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520.
-
Babylon.js feedback on uniformity analysis. gpuweb/gpuweb#2969.
-
Value of reflection attributes if validation fails gpuweb/gpuweb#3064.
-
constrain attribute alignment value to divide device alignment limits gpuweb/gpuweb#2505.
-
Streaming implementations and indirect draws/dispatches gpuweb/gpuweb#2189.
-
[TAG review] Should there be a higher granularity for selecting devices? gpuweb/gpuweb#2945.
-
TAG feedback on WebGPU gpuweb/gpuweb#2927.
-
builtins: reorder to put Overload, Parameterization before Description and notes gpuweb/gpuweb#3044.
-
Questions about index continuity where setBindGroup(index, ...) gpuweb/gpuweb#3042.
-
Open questions about swap chain formats gpuweb/gpuweb#1231.
-
Make error scope objects have same shape as DOMExceptions? gpuweb/gpuweb#1884.
-
Missing depth24unorm format? gpuweb/gpuweb#2732.
-
Specify what happens if stencilClearValue > 255 gpuweb/gpuweb#2921.
-
Update list of attributes for structure members gpuweb/gpuweb#2998.
-
Is there a method similar to "gl.getActiveXXX" in ShaderModule? gpuweb/gpuweb#3003.
-
Shader resources reflection from ShaderModule is needed. gpuweb/gpuweb#2316.
-
maxInterStageShaderComponents is not enough gpuweb/gpuweb#2400.
-
Reintroduce bitshift operators and remove builtins (revert #2713) gpuweb/gpuweb#2986.
-
-
3 issues reopened.
-
Should objects have pointers back to their parents? gpuweb/gpuweb#1497, by @kainino0x.
-
Reading from MSAA depth textures gpuweb/gpuweb#1924, by @kainino0x.
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094, by @kainino0x.
-
-
75 issues received 202 new comments.
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094 (11 comments by @kainino0x, @Kangz, @litherum)
-
Add `saturate()` or `clamp01()` gpuweb/gpuweb#3012 (10 comments by @ben-clayton, @dj2, @dneto0, @jespertheend, @Jiawei-Shao, @kdashg, @magcius)
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766 (9 comments by @greggman, @kainino0x, @Kangz, @kdashg, @sandersdan)
-
Make lost devices appear to function as much as possible gpuweb/gpuweb#1629 (7 comments by @greggman, @kainino0x)
-
Shader stage input-output subsetting rules gpuweb/gpuweb#2038 (7 comments by @jenatali, @kainino0x, @Kangz)
-
WGSL pointer aliasing rules gpuweb/gpuweb#1457 (7 comments by @alan-baker, @kdashg, @litherum, @pow2clk)
-
should WGSL allow shadowing of builtins at module scope gpuweb/gpuweb#2941 (6 comments by @dneto0, @kdashg)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (6 comments by @kainino0x, @Kangz, @r12a, @xfq)
-
Expose WGSL errors in the API gpuweb/gpuweb#2308 (5 comments by @kainino0x, @kdashg, @litherum, @mixtur)
-
wgsl: document how much of an alignment request is checked by the API gpuweb/gpuweb#2999 (5 comments by @dneto0, @kainino0x, @kdashg)
-
Should objects expose the descriptor they were created with? gpuweb/gpuweb#1498 (5 comments by @greggman, @kainino0x, @Kangz)
-
wgsl: Is/should initializer of module-scope variables limited to creation-time expressions? gpuweb/gpuweb#3081 (4 comments by @alan-baker, @ben-clayton, @dneto0)
-
wgsl: comparison operator result is scalar bool when "is scalar" should expand to include abstract numerics. gpuweb/gpuweb#2988 (4 comments by @dneto0, @kdashg)
-
Should objects have pointers back to their parents? gpuweb/gpuweb#1497 (3 comments by @kainino0x, @kdashg, @toji)
-
Consider making ShaderModule compilation errors special gpuweb/gpuweb#2119 (3 comments by @kainino0x, @kdashg, @litherum)
-
wgsl: function param that is pointer-to-workgroup is problematic if the store type contains an atomic gpuweb/gpuweb#3067 (3 comments by @dneto0, @kdashg)
-
Investigation: kill vs demoteToHelper behavior in native APIs gpuweb/gpuweb#921 (3 comments by @dneto0, @kdashg)
-
Investigation: Generate mipmaps gpuweb/gpuweb#386 (3 comments by @devshgraphicsprogramming, @maierfelix)
-
wgsl: Other language types in reserved word list gpuweb/gpuweb#2983 (3 comments by @dneto0, @kdashg, @litherum)
-
Specify a minimum `SizeOf` for a runtime array gpuweb/gpuweb#2997 (3 comments by @kainino0x, @kdashg, @litherum)
-
The validating linear texture data algorithm is overly complex gpuweb/gpuweb#2971 (2 comments by @kainino0x, @Kangz)
-
Include arrays in creation-time constant expressions gpuweb/gpuweb#3056 (2 comments by @dneto0, @kdashg)
-
Clarifications for initializers for creation-time constants gpuweb/gpuweb#3086 (2 comments by @alan-baker, @dj2)
-
[TAG review] Should there be a higher granularity for selecting devices? gpuweb/gpuweb#2945 (2 comments by @kainino0x, @litherum)
-
Streaming implementations and indirect draws/dispatches gpuweb/gpuweb#2189 (2 comments by @Kangz, @litherum)
-
Try-catching getMappedRange is useful but shouldn't be gpuweb/gpuweb#3013 (2 comments by @kainino0x)
-
Is there a method similar to "gl.getActiveXXX" in ShaderModule? gpuweb/gpuweb#3003 (2 comments by @greggman, @Kangz)
-
constrain attribute alignment value to divide device alignment limits gpuweb/gpuweb#2505 (2 comments by @dneto0, @kainino0x)
-
wgsl: expression overloads, and builtin overloads must have all concrete operands or all abstract operands gpuweb/gpuweb#2993 (2 comments by @dj2, @dneto0)
-
vec4<f32> is too hard to type on the keyboard gpuweb/gpuweb#736 (2 comments by @kainino0x, @litherum)
-
List of issues to discuss with TC39, WASM, JS engine teams gpuweb/gpuweb#747 (1 comment by @jimblandy)
-
Criteria for pushing the WebGPU and WGSL specification to Candidate Recommendation gpuweb/gpuweb#3107 (1 comment by @kdashg)
-
wgsl: log accuracy is in terms of an interval of x, not an interval of log(x) gpuweb/gpuweb#3104 (1 comment by @dneto0)
-
Q: will implementations provide additional builtin functions/structures? gpuweb/gpuweb#3102 (1 comment by @Kangz)
-
Consider formatting algorithm calls like function calls() gpuweb/gpuweb#768 (1 comment by @kainino0x)
-
Extend let declaration description to make clear how it is different from const gpuweb/gpuweb#3063 (1 comment by @dj2)
-
Dynamic error description for read-modify-write atomics could use more explanation. gpuweb/gpuweb#3057 (1 comment by @alan-baker)
-
Provide stronger control over physical alignments of buffers and bindings gpuweb/gpuweb#3021 (1 comment by @litherum)
-
[TAG review] Wondering if power consumption choice should be initiated by the user gpuweb/gpuweb#2946 (1 comment by @litherum)
-
Clarify specification around inferring module scoped variable types. gpuweb/gpuweb#3048 (1 comment by @dj2)
-
builtins: reorder to put Overload, Parameterization before Description and notes gpuweb/gpuweb#3044 (1 comment by @dneto0)
-
Make error scope objects have same shape as DOMExceptions? gpuweb/gpuweb#1884 (1 comment by @kainino0x)
-
Separate read-only state for stencil and depth aspects. gpuweb/gpuweb#3036 (1 comment by @kainino0x)
-
Structure types should specify that the contained vectors and matrices must be concretized components gpuweb/gpuweb#3032 (1 comment by @ben-clayton)
-
Specify that structures compare by name, not by structure. gpuweb/gpuweb#2994 (1 comment by @kdashg)
-
GPURenderBundleEncoderDescriptor needs maxCommandCount if it is to be implementable on MTLIndirectCommandBuffer gpuweb/gpuweb#2612 (1 comment by @Kangz)
-
Add query for memory footprint of unsized depth/stencil formats gpuweb/gpuweb#1887 (1 comment by @kainino0x)
-
Reintroduce bitshift operators and remove builtins (revert #2713) gpuweb/gpuweb#2986 (1 comment by @dneto0)
-
61 pull requests opened.
-
Add the "internal" error type gpuweb/gpuweb#3123, by @toji.
-
Allow inter-stage variable subsetting gpuweb/gpuweb#3122, by @kainino0x.
-
Remove styling for multiple <dd>s in a <dl> gpuweb/gpuweb#3121, by @kainino0x.
-
[editorial] Simplify "validating linear texture data" gpuweb/gpuweb#3120, by @kainino0x.
-
Use inequality symbols instead of words in validation rules gpuweb/gpuweb#3119, by @kainino0x.
-
Normalize bikeshed formatting somewhat gpuweb/gpuweb#3118, by @kainino0x.
-
In certain places, identifiers don't resolve to declarations gpuweb/gpuweb#3117, by @dneto0.
-
[editorial] Make GPUCommandBuffer one-time-use gpuweb/gpuweb#3116, by @kainino0x.
-
Restore commas after final members of enums gpuweb/gpuweb#3114, by @kainino0x.
-
Fix abstract numeric cases for comparisons, length builtin gpuweb/gpuweb#3112, by @dneto0.
-
Forbid pointer-to-workgroup param on user functions, when store type … gpuweb/gpuweb#3110, by @dneto0.
-
[test] for #3108 workflow changes gpuweb/gpuweb#3109, by @kainino0x.
-
More workflow rework, fix triggering of build/validate steps gpuweb/gpuweb#3108, by @kainino0x.
-
Add wgsl.lalr.txt to build output for preview.yml workflow gpuweb/gpuweb#3106, by @kainino0x.
-
Fix phrasing of accuracy of cos, sin, log, log2 gpuweb/gpuweb#3105, by @dneto0.
-
Add a note that WebGPU should be included in any high-power warnings gpuweb/gpuweb#3099, by @kainino0x.
-
"Fake" transfers of getMappedRange ArrayBuffers gpuweb/gpuweb#3098, by @kainino0x.
-
nit: remove a redundant validation step gpuweb/gpuweb#3097, by @kainino0x.
-
Update WGSL README, to discuss LALR(1) validation gpuweb/gpuweb#3096, by @dneto0.
-
Fix definition of matrix-column-vector product. gpuweb/gpuweb#3094, by @jimblandy.
-
Correspondence References: Resources gpuweb/gpuweb#3093, by @kainino0x.
-
Add "Correspondence Reference" stub doc gpuweb/gpuweb#3092, by @kainino0x.
-
Test change for GitHub Actions configuration gpuweb/gpuweb#3091, by @kainino0x.
-
Clarify color space section, add issue gpuweb/gpuweb#3090, by @kainino0x.
-
Allow drawCount == maxDrawCount gpuweb/gpuweb#3089, by @takahirox.
-
Correspondence Reference: Limits gpuweb/gpuweb#3088, by @kainino0x.
-
Stop using spec-prod for gh-pages builds gpuweb/gpuweb#3087, by @kainino0x.
-
Set a limit for maximum array size gpuweb/gpuweb#3078, by @alan-baker.
-
Allow inter-stage variable subsetting by prefix gpuweb/gpuweb#3073, by @kainino0x.
-
Allow BGL entries for multisampled:true "depth" textures gpuweb/gpuweb#3072, by @kainino0x.
-
Replace section numbers with section anchor names gpuweb/gpuweb#3071, by @kainino0x.
-
Allow depth textures with cube/array/multisampled types gpuweb/gpuweb#3070, by @kainino0x.
-
[editorial] Generalize language of reading-depth-stencil section gpuweb/gpuweb#3069, by @kainino0x.
-
Add saturate floating point builtin function gpuweb/gpuweb#3068, by @dneto0.
-
fix bikeshed syntax typo gpuweb/gpuweb#3066, by @kainino0x.
-
Ensure that reflection attributes are initialized gpuweb/gpuweb#3065, by @toji.
-
wgsl: Fix description of atomicCompareExchangeWeak gpuweb/gpuweb#3055, by @jrprice.
-
Use a responsive SVG for the logo. gpuweb/gpuweb#3054, by @greggman.
-
Correct '/' to '.' in 'derive render targets layout from pipeline'. gpuweb/gpuweb#3041, by @jimblandy.
-
Clean methods, separate build, and apply suggestion for WGSL grammar gpuweb/gpuweb#3040, by @mehmetoguzderin.
-
editorial: specify units for limits where relevant gpuweb/gpuweb#3039, by @kainino0x.
-
Allow using depth textures with texture_2d gpuweb/gpuweb#3038, by @kainino0x.
-
Add WGSL strict LALR(1) check to CI gpuweb/gpuweb#3037, by @dneto0.
-
Add limit for max total render target size per pixel gpuweb/gpuweb#3028, by @kainino0x.
-
Combine section headers which define the same function gpuweb/gpuweb#3020, by @dj2.
-
Align phrasing of Firing an event with DOM gpuweb/gpuweb#3019, by @dontcallmedom.
-
Ensure the formats supporting multisampling must support rendering gpuweb/gpuweb#3018, by @Jiawei-Shao.
-
Move builtin names to section headers. gpuweb/gpuweb#3016, by @dj2.
-
Expose the buffer state as an attribute gpuweb/gpuweb#3014, by @kainino0x.
-
Format internal slots consistently with IDL attributes gpuweb/gpuweb#3011, by @kainino0x.
-
Return null from getMappedRange if mapping is pending gpuweb/gpuweb#3010, by @kainino0x.
-
Even more descriptor definitions gpuweb/gpuweb#3009, by @toji.
-
Reformat builtin tables. gpuweb/gpuweb#3008, by @dj2.
-
Update WGSL stage syntax in example code gpuweb/gpuweb#3006, by @toji.
-
Add GPURenderPassDescriptor.maxDrawCount gpuweb/gpuweb#3005, by @Kangz.
-
Specify minimum permitted values of alignment limits gpuweb/gpuweb#3002, by @kainino0x.
-
Add interpolate, invariant to attribs that can be applied to struct m… gpuweb/gpuweb#3000, by @dneto0.
-
ldexp: all operands concrete or all operands abstract gpuweb/gpuweb#2995, by @dneto0.
-
Fix typos gpuweb/gpuweb#2992, by @beaufortfrancois.
-
Fix typo gpuweb/gpuweb#2991, by @beaufortfrancois.
-
Add WebGPU Adapter Info Registry document gpuweb/gpuweb#2990, by @beaufortfrancois.
-
-
59 pull requests merged.
-
[editorial] Simplify "validating linear texture data" gpuweb/gpuweb#3120.
-
Remove styling for multiple <dd>s in a <dl> gpuweb/gpuweb#3121.
-
Use inequality symbols instead of words in validation rules gpuweb/gpuweb#3119.
-
Normalize bikeshed formatting somewhat gpuweb/gpuweb#3118.
-
[editorial] Make GPUCommandBuffer one-time-use gpuweb/gpuweb#3116.
-
Forbid pointer-to-workgroup param on user functions, when store type … gpuweb/gpuweb#3110.
-
Fix abstract numeric cases for comparisons, length builtin gpuweb/gpuweb#3112.
-
More workflow rework, fix triggering of build/validate steps gpuweb/gpuweb#3108.
-
Add wgsl.lalr.txt to build output for preview.yml workflow gpuweb/gpuweb#3106.
-
Add "Correspondence Reference" stub doc gpuweb/gpuweb#3092.
-
Fix phrasing of accuracy of cos, sin, log, log2 gpuweb/gpuweb#3105.
-
Update WGSL README, to discuss LALR(1) validation gpuweb/gpuweb#3096.
-
"Fake" transfers of getMappedRange ArrayBuffers gpuweb/gpuweb#3098.
-
nit: remove a redundant validation step gpuweb/gpuweb#3097.
-
Add a note that WebGPU should be included in any high-power warnings gpuweb/gpuweb#3099.
-
Update WGSL README: instructions about tree-sitter gpuweb/gpuweb#2126.
-
WGSL: Explain parameterized type rules more explicitly. gpuweb/gpuweb#2682.
-
Fix definition of matrix-column-vector product. gpuweb/gpuweb#3094.
-
Allow drawCount == maxDrawCount gpuweb/gpuweb#3089.
-
Stop using spec-prod for gh-pages builds gpuweb/gpuweb#3087.
-
Clarify color space section, add issue gpuweb/gpuweb#3090.
-
[editorial] Generalize language of reading-depth-stencil section gpuweb/gpuweb#3069.
-
fix bikeshed syntax typo gpuweb/gpuweb#3066.
-
Ensure that reflection attributes are initialized gpuweb/gpuweb#3065.
-
Allow using depth textures with texture_2d gpuweb/gpuweb#3038.
-
Specify minimum permitted values of alignment limits gpuweb/gpuweb#3002.
-
Add GPURenderPassDescriptor.maxDrawCount gpuweb/gpuweb#3005.
-
wgsl: Fix description of atomicCompareExchangeWeak gpuweb/gpuweb#3055.
-
Use a responsive SVG for the logo. gpuweb/gpuweb#3054.
-
Correct '/' to '.' in 'derive render targets layout from pipeline'. gpuweb/gpuweb#3041.
-
Combine section headers which define the same function gpuweb/gpuweb#3020.
-
Clean methods, separate build, and apply suggestion for WGSL grammar gpuweb/gpuweb#3040.
-
Add WGSL strict LALR(1) check to CI gpuweb/gpuweb#3037.
-
editorial: specify units for limits where relevant gpuweb/gpuweb#3039.
-
Add WGSL grammar analyzer gpuweb/gpuweb#2754.
-
Ensure the formats supporting multisampling must support rendering gpuweb/gpuweb#3018.
-
Align phrasing of Firing an event with DOM gpuweb/gpuweb#3019.
-
Move builtin names to section headers. gpuweb/gpuweb#3016.
-
wgsl: Allow phony assignment of abstract numeric type gpuweb/gpuweb#2968.
-
Add maxColorAttachments missing from GPUSupportedLimits gpuweb/gpuweb#2985.
-
Support building multi-file specs with the Bikeshed API gpuweb/gpuweb#2856.
-
Format internal slots consistently with IDL attributes gpuweb/gpuweb#3011.
-
Reformat builtin tables. gpuweb/gpuweb#3008.
-
Use table formatting for texture builtins. gpuweb/gpuweb#2906.
-
Use PredefinedColorSpace instead of GPUPredefinedColorSpace gpuweb/gpuweb#2967.
-
Remove depth24unorm-stencil8 gpuweb/gpuweb#2950.
-
Clarify to interpret stencilClearValue as masked gpuweb/gpuweb#2964.
-
Add reflection for container objects. gpuweb/gpuweb#2919.
-
Even more descriptor definitions gpuweb/gpuweb#3009.
-
Update WGSL stage syntax in example code gpuweb/gpuweb#3006.
-
Add interpolate, invariant to attribs that can be applied to struct m… gpuweb/gpuweb#3000.
-
Add maxInterStageShaderVariables and inter-stage limit validation gpuweb/gpuweb#2857.
-
Fix typos gpuweb/gpuweb#2992.
-
ldexp: all operands concrete or all operands abstract gpuweb/gpuweb#2995.
-
Explicitly state the accuracy for clamp gpuweb/gpuweb#2918.
-
[wgsl]: Clarify attribute Valid Values. gpuweb/gpuweb#2924.
-
wgsl: Use code tags for enable f16 code snippet. gpuweb/gpuweb#2974.
-
Add shift operators, remove shiftLeft, shiftRight builtins gpuweb/gpuweb#2989.
-
Fix typo gpuweb/gpuweb#2991.
-
-
12 pull requests closed.
-
Allow inter-stage variable subsetting by prefix gpuweb/gpuweb#3073.
-
Replace section numbers with section anchor names gpuweb/gpuweb#3071.
-
[test] for #3108 workflow changes gpuweb/gpuweb#3109.
-
Add limit for max total render target size per pixel gpuweb/gpuweb#3028.
-
Rewrite explanation of number tokens gpuweb/gpuweb#1454.
-
Resolve numeric literal types through inference. gpuweb/gpuweb#2651.
-
Test change for GitHub Actions configuration gpuweb/gpuweb#3091.
-
Rename `discard` to `terminate`. gpuweb/gpuweb#2948.
-
Return null from getMappedRange if mapping is pending gpuweb/gpuweb#3010.
-
Mirror descriptor properties on GPU resources gpuweb/gpuweb#2494.
-
Explicitly state the accuracy for ldexp gpuweb/gpuweb#2917.
-
Set the default value of maxInterStageShaderComponents to 32 gpuweb/gpuweb#2763.
-
-
49 pull requests received 185 new comments.
-
Add shift operators, remove shiftLeft, shiftRight builtins gpuweb/gpuweb#2989 (26 comments by @alan-baker, @dj2, @dneto0)
-
Clean methods, separate build, and apply suggestion for WGSL grammar gpuweb/gpuweb#3040 (12 comments by @kainino0x, @mehmetoguzderin)
-
Set a limit for maximum array size gpuweb/gpuweb#3078 (7 comments by @alan-baker, @ben-clayton, @dneto0, @kdashg)
-
Support building multi-file specs with the Bikeshed API gpuweb/gpuweb#2856 (6 comments by @kainino0x, @mehmetoguzderin)
-
Correspondence Reference: Limits gpuweb/gpuweb#3088 (5 comments by @cwfitzgerald, @kainino0x, @toji)
-
Add GPURenderPassDescriptor.maxDrawCount gpuweb/gpuweb#3005 (5 comments by @kainino0x, @Kangz, @litherum, @takahirox)
-
Ensure the formats supporting multisampling must support rendering gpuweb/gpuweb#3018 (5 comments by @kainino0x, @Kangz)
-
Allow inter-stage variable subsetting by prefix gpuweb/gpuweb#3073 (4 comments by @kainino0x, @Kangz)
-
Add saturate floating point builtin function gpuweb/gpuweb#3068 (4 comments by @alan-baker, @dneto0)
-
Add WebGPU Adapter Info Registry document gpuweb/gpuweb#2990 (4 comments by @kainino0x, @tidoust, @toji)
-
Return null from getMappedRange if mapping is pending gpuweb/gpuweb#3010 (4 comments by @kainino0x, @Kangz)
-
Add maxInterStageShaderVariables and inter-stage limit validation gpuweb/gpuweb#2857 (4 comments by @Jiawei-Shao, @kainino0x)
-
In certain places, identifiers don't resolve to declarations gpuweb/gpuweb#3117 (3 comments by @alan-baker, @dneto0)
-
Add limit for max total render target size per pixel gpuweb/gpuweb#3028 (3 comments by @kainino0x, @Kangz, @litherum)
-
Use inequality symbols instead of words in validation rules gpuweb/gpuweb#3119 (2 comments by @kainino0x, @toji)
-
Update WGSL README: instructions about tree-sitter gpuweb/gpuweb#2126 (2 comments by @dneto0, @mehmetoguzderin)
-
Stop using spec-prod for gh-pages builds gpuweb/gpuweb#3087 (2 comments by @kainino0x, @mehmetoguzderin)
-
wgsl: Allow phony assignment of abstract numeric type gpuweb/gpuweb#2968 (2 comments by @ben-clayton, @kdashg)
-
Add proposals/ directory and pipeline statistics extension doc gpuweb/gpuweb#2301 (1 comment by @kainino0x)
-
Use PredefinedColorSpace instead of GPUPredefinedColorSpace gpuweb/gpuweb#2967 (1 comment by @kainino0x)
-
Add interpolate, invariant to attribs that can be applied to struct m… gpuweb/gpuweb#3000 (1 comment by @dneto0)
-
TR publications
-
36 documents published.
-
WebGPU published 18 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 18 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/gpuweb
-
82 issues opened.
-
Reintroduce bitshift operators and remove builtins (revert #2713) gpuweb/gpuweb#2986, by @kdashg.
-
wgsl: Other language types in reserved word list gpuweb/gpuweb#2983, by @dj2.
-
wgsl: Mark type constructors `@const`? gpuweb/gpuweb#2982, by @dj2.
-
wgsl: The term `override expression` is confusing gpuweb/gpuweb#2981, by @dj2.
-
wgsl: `storageBarrier` memory scope question gpuweb/gpuweb#2980, by @dj2.
-
atomic types section references `QueueFamily` which doesn't seem to be defined gpuweb/gpuweb#2979, by @dj2.
-
wgsl: atomic type section should link to memory model section gpuweb/gpuweb#2978, by @dj2.
-
wgsl: Atomic Types should mention initialization gpuweb/gpuweb#2977, by @dj2.
-
wgsl: key use cases for Vector and Matrix seem out of place gpuweb/gpuweb#2976, by @dj2.
-
wgsl: Add some clarification in the `Scalar Types` section that an Abstract is not a scalar type gpuweb/gpuweb#2975, by @dj2.
-
Boolean Type rules table still references SPIR-V gpuweb/gpuweb#2973, by @dj2.
-
Reorder `Types` section for clarity gpuweb/gpuweb#2972, by @dj2.
-
The validating linear texture data algorithm is overly complex gpuweb/gpuweb#2971, by @Kangz.
-
Clarify meaning of `context` in integer example in `TYPE INFERENCE FOR LITERALS` gpuweb/gpuweb#2970, by @dj2.
-
Babylon.js feedback on uniformity analysis. gpuweb/gpuweb#2969, by @Kangz.
-
maxColorAttachments not sufficient? gpuweb/gpuweb#2965, by @greggman.
-
Distinction between content-timeline and device-timeline state gpuweb/gpuweb#2962, by @kainino0x.
-
Improved style for issuing work between timelines gpuweb/gpuweb#2961, by @kainino0x.
-
Conversion expression preamble claims no conversion or promotion of numerics. gpuweb/gpuweb#2960, by @dj2.
-
Conversion expressions for the Abstract types gpuweb/gpuweb#2959, by @dj2.
-
Split spec into multiple files and structure it consistently gpuweb/gpuweb#2958, by @kainino0x.
-
Consider how to make the spec a better developer reference gpuweb/gpuweb#2957, by @kainino0x.
-
Make return types for `modf` and `frexp` generic gpuweb/gpuweb#2955, by @dj2.
-
AbstractFloat should not be convertible to infinity or NaN in concrete float type gpuweb/gpuweb#2953, by @dneto0.
-
Build is broken gpuweb/gpuweb#2951, by @kainino0x.
-
[TAG review] Have a note suggesting surface a warning about power usage? gpuweb/gpuweb#2947, by @kdashg.
-
[TAG review] Wondering if power consumption choice should be initiated by the user gpuweb/gpuweb#2946, by @kdashg.
-
[TAG review] Should there be a higher granularity for selecting devices? gpuweb/gpuweb#2945, by @kdashg.
-
wgsl editorial: change ref<S,T,A> -> ref<A,T,M> gpuweb/gpuweb#2943, by @dneto0.
-
should WGSL allow shadowing of builtins at module scope gpuweb/gpuweb#2941, by @dneto0.
-
Specify the meaning of `S` in the ConversionRank table gpuweb/gpuweb#2940, by @dj2.
-
Update note about when automatic conversions occur. gpuweb/gpuweb#2939, by @dj2.
-
Example of unary negation overloads is out of date. gpuweb/gpuweb#2938, by @dj2.
-
Update or remove note about function declarations containing other declarations gpuweb/gpuweb#2936, by @dj2.
-
Valid and Invalid declarations comment for built-in functions is a bit misleading gpuweb/gpuweb#2934, by @dj2.
-
Add example of shadowing function parameter in `Valid and invalid declarations` example gpuweb/gpuweb#2933, by @dj2.
-
Spec prose seems to allow `staticAssert` before `global_directive` gpuweb/gpuweb#2931, by @dj2.
-
Review HTML integration and more tightly integrate for resizing gpuweb/gpuweb#2930, by @annevk.
-
Empty vertex buffer slots should be skipped for OOB validation in draw calls gpuweb/gpuweb#2929, by @takahirox.
-
TAG feedback on WebGPU gpuweb/gpuweb#2927, by @Kangz.
-
importExternalTexture crashes the browser gpuweb/gpuweb#2923, by @gaoyakun.
-
Specify what happens if stencilClearValue > 255 gpuweb/gpuweb#2921, by @Kangz.
-
editorial bugs in WGSL spec gpuweb/gpuweb#2920, by @dneto0.
-
Clarify `Valid Values` language in the attributes table. gpuweb/gpuweb#2916, by @dj2.
-
`workgroup_size` valid values gpuweb/gpuweb#2915, by @dj2.
-
wgsl: write examples of overload resolution for switch gpuweb/gpuweb#2914, by @dneto0.
-
`@interpolate` values on reserved word list gpuweb/gpuweb#2913, by @dj2.
-
Confusing languages in technical overview around conversions. gpuweb/gpuweb#2910, by @dj2.
-
`textureDimension` missing `i32` overload for `level` parameter gpuweb/gpuweb#2908, by @dj2.
-
Clarify the normalized float ranges in the `pack` methods gpuweb/gpuweb#2903, by @dj2.
-
textureSampleLevel missing `external` in definition of `t` gpuweb/gpuweb#2896, by @dj2.
-
`textureSampleCompare` lists `texture_depth_cube_array` twice. gpuweb/gpuweb#2895, by @dj2.
-
`external` textures and `textureSample` gpuweb/gpuweb#2894, by @dj2.
-
Add texture types to `t` for textureSampleGrad and textureSampleBias gpuweb/gpuweb#2892, by @dj2.
-
textureSampleCompareLevel missing params and return documentation. gpuweb/gpuweb#2891, by @dj2.
-
wgsl: numeric params of attributes should be creation-time expressions not literals gpuweb/gpuweb#2890, by @dneto0.
-
Defining "Undefined Value" gpuweb/gpuweb#2888, by @kdashg.
-
forbid spelling out pointer types that are not instantiable gpuweb/gpuweb#2886, by @dneto0.
-
Precision when behavior of float built-in functions is implementation-specific gpuweb/gpuweb#2884, by @ghadeeras.
-
editorial: define scalar type categories: signed int, unsigned int, floating gpuweb/gpuweb#2883, by @dneto0.
-
`textureNumLayers` lists `multisampled` as a possible `t` value gpuweb/gpuweb#2881, by @dj2.
-
textureLoad does not specify bounds for sample_index gpuweb/gpuweb#2879, by @dj2.
-
Render to 3D texture slices gpuweb/gpuweb#2877, by @brendan-duncan.
-
Behaviour of overflowing arithmetic operations gpuweb/gpuweb#2875, by @hasali19.
-
Error handling in *beginRenderPass* and *beginComputePass* is underspecified gpuweb/gpuweb#2874, by @jimblandy.
-
editorial: add example showing how to write an i32 of most negative value gpuweb/gpuweb#2872, by @dneto0.
-
Initialization of `GPUObjectBase` internal slot is left unspecified gpuweb/gpuweb#2871, by @jimblandy.
-
Consider "magic" format that can be either rgba8unorm or bgra8unorm gpuweb/gpuweb#2869, by @kainino0x.
-
Should cullMode require "none" for points and lines? gpuweb/gpuweb#2863, by @toji.
-
Uniformity analysis of local variables gpuweb/gpuweb#2859, by @alan-baker.
-
`textureGather` limits on `coords` and `array_index` gpuweb/gpuweb#2850, by @dj2.
-
Clarification around `sampled type` in the Texture Builtin section. gpuweb/gpuweb#2849, by @dj2.
-
maxComputeWorkgroupStorageSize limit gpuweb/gpuweb#2832, by @jereroy.
-
Referencing the Unicode Standard gpuweb/gpuweb#2829, by @xfq.
-
WGSL identifier comparison gpuweb/gpuweb#2828, by @xfq.
-
Clarify behaviour of `fract` with very small negative numbers gpuweb/gpuweb#2822, by @zoddicus.
-
Parameterization missing for `arrayLength` gpuweb/gpuweb#2818, by @dj2.
-
Provide a description of every descriptor member gpuweb/gpuweb#2815, by @toji.
-
69 issues closed.
-
Are padding bytes of structures intialized with zero value expressions? gpuweb/gpuweb#1747.
-
WGSL parsing assumes context-aware tokenization gpuweb/gpuweb#2717.
-
wgsl editorial: change ref<S,T,A> -> ref<A,T,M> gpuweb/gpuweb#2943.
-
Build is broken gpuweb/gpuweb#2951.
-
Empty vertex buffer slots should be skipped for OOB validation in draw calls gpuweb/gpuweb#2929.
-
Does compositingAlphaMode affect using the canvas as an image source? gpuweb/gpuweb#1847.
-
WebGPURenderingContext should support preserveDrawingBuffer: true gpuweb/gpuweb#2743.
-
Specify the meaning of `S` in the ConversionRank table gpuweb/gpuweb#2940.
-
[WGSL] WGSL needs a plan for expansion gpuweb/gpuweb#600.
-
Should the spec mention location aliasing? gpuweb/gpuweb#956.
-
☂️ Initial internal feedback gpuweb/gpuweb#2213.
-
importExternalTexture crashes the browser gpuweb/gpuweb#2923.
-
GPUExternalTexture [[active_imports]] only works if there are RAFs gpuweb/gpuweb#2785.
-
Metal primitive topology required for the shader compilation hint gpuweb/gpuweb#2574.
-
Zero initializing textures can not be done efficiently gpuweb/gpuweb#1422.
-
Missing specification of error type produced gpuweb/gpuweb#2730.
-
rgba8unorm canvas textures and IOSurface gpuweb/gpuweb#2535.
-
Behavior of discarding everything gpuweb/gpuweb#2523.
-
forbid spelling out pointer types that are not instantiable gpuweb/gpuweb#2886.
-
Attribute 'Valid Values' should be updated for abstract integers gpuweb/gpuweb#2809.
-
Spec clarification around untyped LHS and abstract RHS gpuweb/gpuweb#2802.
-
Clarify behaviour of `fract` with very small negative numbers gpuweb/gpuweb#2822.
-
Accuracy of dot built-in never specified gpuweb/gpuweb#2808.
-
Should GPUDevice.destroy detach all buffer mapping ArrayBuffers? gpuweb/gpuweb#2545.
-
createRenderPipeline with preferred format is dependent on canvas gpuweb/gpuweb#2747.
-
`textureSampleCompare` lists `texture_depth_cube_array` twice. gpuweb/gpuweb#2895.
-
Should cullMode require "none" for points and lines? gpuweb/gpuweb#2863.
-
Number literal conversion clarification gpuweb/gpuweb#2769.
-
Explicit flat interpolation requirement clarification gpuweb/gpuweb#2760.
-
WGSL identifier comparison gpuweb/gpuweb#2828.
-
`textureNumLayers` lists `multisampled` as a possible `t` value gpuweb/gpuweb#2881.
-
textureLoad does not specify bounds for sample_index gpuweb/gpuweb#2879.
-
Error handling in *beginRenderPass* and *beginComputePass* is underspecified gpuweb/gpuweb#2874.
-
Short i18n review for WebGPU / WGSL gpuweb/gpuweb#2546.
-
The preferred format is not always storage-capable gpuweb/gpuweb#2748.
-
C-style casts gpuweb/gpuweb#2220.
-
Clarify atomics initialization for variables in workgroup address space gpuweb/gpuweb#2811.
-
Spec guidance on when/where error messages appear gpuweb/gpuweb#2516.
-
`textureGather` limits on `coords` and `array_index` gpuweb/gpuweb#2850.
-
Determine whether getBindGroupLayout returns a new object or not gpuweb/gpuweb#2501.
-
wgsl: Missing specification on modf behavior for negatives. gpuweb/gpuweb#2468.
-
Consider making `binding_array` a reserved word gpuweb/gpuweb#2794.
-
Consider copyToTexture interactions with high-bit-depth source images gpuweb/gpuweb#1482.
-
GPUTextureFormatDescriptor.viewFormats performance hint is nonhelpful gpuweb/gpuweb#2573.
-
Are duplicate timestamp locations allowed on the same pass? gpuweb/gpuweb#2662.
-
What do the GPUBufferUsageFlags constants mean? gpuweb/gpuweb#2663.
-
Static assert gpuweb/gpuweb#2584.
-
Third param of `mix` with auto-splat param should be `T`. gpuweb/gpuweb#2806.
-
Return type of second form of `mix` seems incorrect. gpuweb/gpuweb#2807.
-
`@const` missing on some built-in functions gpuweb/gpuweb#2787.
-
Remove serialization/transfer from the spec gpuweb/gpuweb#2629.
-
Make automatic bind group layouts explicit gpuweb/gpuweb#2636.
-
specify behaviour of floating point operations on NaNs gpuweb/gpuweb#1711.
-
WGSL grammar needs update for var type inference gpuweb/gpuweb#1805.
-
WGSL: Allow global let-declarations initialized with constant expressions gpuweb/gpuweb#1911.
-
shader programming model: use little-endian layouts for scalar data gpuweb/gpuweb#234.
-
Post-comparison-filterable depth formats gpuweb/gpuweb#1266.
-
Integer signedness consistency gpuweb/gpuweb#2570.
-
Relax the out-of-bounds specification gpuweb/gpuweb#1649.
-
114 issues received 316 new comments.
-
Should we constrain the location of user input-output stage variables WGSL gpuweb/gpuweb#1962 (22 comments by @Jiawei-Shao, @kainino0x, @Kangz)
-
WebGPURenderingContext should support preserveDrawingBuffer: true gpuweb/gpuweb#2743 (18 comments by @greggman, @kainino0x, @kdashg, @litherum)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (13 comments by @greggman, @kainino0x, @Kangz, @mehmetoguzderin, @xfq)
-
should WGSL allow shadowing of builtins at module scope gpuweb/gpuweb#2941 (11 comments by @ben-clayton, @dj2, @dneto0, @kainino0x)
-
Clarify behavior of float division/remainder by 0 gpuweb/gpuweb#2798 (9 comments by @jiangzhaoming, @kdashg, @m-schuetz, @teoxoy)
-
Relax the null buffer check for zero-attribute vertex buffers in "valid to draw with GPURenderCommandsMixin" gpuweb/gpuweb#2864 (8 comments by @kainino0x, @Kangz, @takahirox)
-
Are padding bytes of structures intialized with zero value expressions? gpuweb/gpuweb#1747 (7 comments by @alan-baker, @kdashg, @litherum)
-
vec4<f32> is too hard to type on the keyboard gpuweb/gpuweb#736 (6 comments by @ben-clayton, @kainino0x, @kdashg, @litherum)
-
Metal primitive topology required for the shader compilation hint gpuweb/gpuweb#2574 (6 comments by @cwfitzgerald, @jimblandy, @kainino0x, @Kangz, @kvark)
-
WGSL should forbid pointers to runtime-sized values as user-defined function parameters gpuweb/gpuweb#2268 (5 comments by @dneto0, @jimblandy, @kdashg)
-
wgsl: Missing specification on modf behavior for negatives. gpuweb/gpuweb#2468 (5 comments by @dneto0, @kdashg)
-
AbstractFloat should not be convertible to infinity or NaN in concrete float type gpuweb/gpuweb#2953 (4 comments by @ben-clayton, @dneto0, @kdashg)
-
Specify what happens if stencilClearValue > 255 gpuweb/gpuweb#2921 (4 comments by @Jiawei-Shao, @kainino0x)
-
Clarify `Valid Values` language in the attributes table. gpuweb/gpuweb#2916 (4 comments by @dj2, @dneto0)
-
Spec clarification around untyped LHS and abstract RHS gpuweb/gpuweb#2802 (4 comments by @dj2, @dneto0, @kdashg)
-
Attribute 'Valid Values' should be updated for abstract integers gpuweb/gpuweb#2809 (4 comments by @alan-baker, @dneto0, @kdashg)
-
Clarify behaviour of `fract` with very small negative numbers gpuweb/gpuweb#2822 (4 comments by @dneto0, @greggman, @kdashg, @zoddicus)
-
GPUBuffer.mapAsync is overly pessimistic gpuweb/gpuweb#2646 (4 comments by @kainino0x, @Kangz, @litherum)
-
Precision when behavior of float built-in functions is implementation-specific gpuweb/gpuweb#2884 (4 comments by @alan-baker, @ghadeeras, @kdashg)
-
Referencing the Unicode Standard gpuweb/gpuweb#2829 (4 comments by @dneto0, @kdashg, @mehmetoguzderin)
-
`textureGather` limits on `coords` and `array_index` gpuweb/gpuweb#2850 (4 comments by @alan-baker, @dj2)
-
wgsl: Mark type constructors `@const`? gpuweb/gpuweb#2982 (3 comments by @alan-baker, @dj2, @teoxoy)
-
wgsl: Atomic Types should mention initialization gpuweb/gpuweb#2977 (3 comments by @alan-baker, @dj2)
-
Investigation: kill vs demoteToHelper behavior in native APIs gpuweb/gpuweb#921 (3 comments by @kdashg, @litherum)
-
Combined limit for max vertex stage bind groups + vertex buffer bindings gpuweb/gpuweb#2749 (3 comments by @kainino0x, @litherum)
-
GPUExternalTexture [[active_imports]] only works if there are RAFs gpuweb/gpuweb#2785 (3 comments by @kainino0x, @Kangz)
-
constrain attribute alignment value to divide device alignment limits gpuweb/gpuweb#2505 (3 comments by @kainino0x, @kdashg)
-
[wgsl] Encoding for shader text needs to be specified gpuweb/gpuweb#565 (2 comments by @kdashg, @mehmetoguzderin)
-
atomic types section references `QueueFamily` which doesn't seem to be defined gpuweb/gpuweb#2979 (2 comments by @alan-baker, @dj2)
-
context-aware tokenization: Sometimes >> is better parsed as two copies of > (similar for ]] ) gpuweb/gpuweb#2092 (2 comments by @kdashg)
-
Confusing languages in technical overview around conversions. gpuweb/gpuweb#2910 (2 comments by @dneto0, @kdashg)
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520 (2 comments by @kainino0x)
-
How should WGSL map to the availability/visibility parts of the Vulkan Memory Model? gpuweb/gpuweb#2228 (2 comments by @kdashg)
-
Does compositingAlphaMode affect using the canvas as an image source? gpuweb/gpuweb#1847 (2 comments by @kainino0x, @shaoboyan)
-
wgsl: numeric params of attributes should be creation-time expressions not literals gpuweb/gpuweb#2890 (2 comments by @alan-baker, @dneto0)
-
Error handling in *beginRenderPass* and *beginComputePass* is underspecified gpuweb/gpuweb#2874 (2 comments by @kainino0x, @toji)
-
WGSL grammar needs update for var type inference gpuweb/gpuweb#1805 (2 comments by @jrprice, @kdashg)
-
wgsl: comparison operator result is scalar bool when "is scalar" should expand to include abstract numerics. gpuweb/gpuweb#2988 (1 comment by @dneto0)
-
Uniformity analysis: find a way to load from a storage buffer uniformly. gpuweb/gpuweb#2321 (1 comment by @kdashg)
-
The validating linear texture data algorithm is overly complex gpuweb/gpuweb#2971 (1 comment by @Kangz)
-
Potential parsing conflict with const literals and postfix expressions. gpuweb/gpuweb#2772 (1 comment by @dneto0)
-
Distinction between content-timeline and device-timeline state gpuweb/gpuweb#2962 (1 comment by @kainino0x)
-
[TAG review] Wondering if power consumption choice should be initiated by the user gpuweb/gpuweb#2946 (1 comment by @greggman)
-
[TAG review] Have a note suggesting surface a warning about power usage? gpuweb/gpuweb#2947 (1 comment by @greggman)
-
Spec prose seems to allow `staticAssert` before `global_directive` gpuweb/gpuweb#2931 (1 comment by @alan-baker)
-
Clarification around `sampled type` in the Texture Builtin section. gpuweb/gpuweb#2849 (1 comment by @kdashg)
-
forbid spelling out pointer types that are not instantiable gpuweb/gpuweb#2886 (1 comment by @kdashg)
-
`textureSampleCompare` lists `texture_depth_cube_array` twice. gpuweb/gpuweb#2895 (1 comment by @dj2)
-
column_major and row_major matrix annotations, WRT pre/post matrix multiplication gpuweb/gpuweb#2481 (1 comment by @kdashg)
-
createRenderPipeline with preferred format is dependent on canvas gpuweb/gpuweb#2747 (1 comment by @kainino0x)
-
wgsl: add 'array' overload to construct an array from components, infering type and count gpuweb/gpuweb#2346 (1 comment by @dneto0)
-
Determine whether getBindGroupLayout returns a new object or not gpuweb/gpuweb#2501 (1 comment by @kainino0x)
-
Clarify the buffer slots validation rule in "valid to draw with GPURenderCommandsMixin" gpuweb/gpuweb#2843 (1 comment by @takahirox)
-
WGSL: Allow global let-declarations initialized with constant expressions gpuweb/gpuweb#1911 (1 comment by @kdashg)
-
shader programming model: use little-endian layouts for scalar data gpuweb/gpuweb#234 (1 comment by @kdashg)
-
-
92 pull requests opened.
-
Add shift operators, remove shiftLeft, shiftRight builtins gpuweb/gpuweb#2989, by @dneto0.
-
Disallow accessing padding gpuweb/gpuweb#2987, by @alan-baker.
-
Add maxColorAttachments missing from GPUSupportedLimits gpuweb/gpuweb#2985, by @kainino0x.
-
AbstractFloat to out-of-finite-range is shader creation error gpuweb/gpuweb#2984, by @dneto0.
-
wgsl: Use code tags for enable f16 code snippet. gpuweb/gpuweb#2974, by @dj2.
-
wgsl: Allow phony assignment of abstract numeric type gpuweb/gpuweb#2968, by @ben-clayton.
-
Use PredefinedColorSpace instead of GPUPredefinedColorSpace gpuweb/gpuweb#2967, by @Kangz.
-
Remove commas after final member of a WebIDL enum. gpuweb/gpuweb#2966, by @jimblandy.
-
Clarify to interpret stencilClearValue as masked gpuweb/gpuweb#2964, by @Jiawei-Shao.
-
Fix proposed IDL in adapters identifier explainer gpuweb/gpuweb#2963, by @beaufortfrancois.
-
Better specify color space conversion elision gpuweb/gpuweb#2956, by @kainino0x.
-
'utilize' -> 'use' gpuweb/gpuweb#2954, by @dneto0.
-
Fix and clean biblio references gpuweb/gpuweb#2952, by @tidoust.
-
Remove depth24unorm-stencil8 gpuweb/gpuweb#2950, by @kainino0x.
-
Editorial nits in depth24plus section gpuweb/gpuweb#2949, by @kainino0x.
-
Rename `discard` to `terminate`. gpuweb/gpuweb#2948, by @kdashg.
-
Change ref<S,T,A> to ref<AS,T,AM> gpuweb/gpuweb#2944, by @dneto0.
-
Clarify 'S' in the ref<S,T,A> row of the ConversionRank gpuweb/gpuweb#2942, by @dneto0.
-
[wgsl]: Minor grammar cleanup, add spec link. gpuweb/gpuweb#2937, by @dj2.
-
OOB check in draw calls should ignore empty vertex buffer slots gpuweb/gpuweb#2935, by @takahirox.
-
Fix cross-reference to pointer type gpuweb/gpuweb#2932, by @dneto0.
-
Specify the meaning of `Foo.Interface`. gpuweb/gpuweb#2928, by @jimblandy.
-
mapAsync(WRITE) and zero-filling gpuweb/gpuweb#2926, by @litherum.
-
[wgsl]: Remove interpolation tokens from reserved words gpuweb/gpuweb#2925, by @dj2.
-
[wgsl]: Clarify attribute Valid Values. gpuweb/gpuweb#2924, by @dj2.
-
Correct sense of validation test in GPURenderBundleEncoder.finish. gpuweb/gpuweb#2922, by @jimblandy.
-
Add reflection for container objects. gpuweb/gpuweb#2919, by @Kangz.
-
Explicitly state the accuracy for clamp gpuweb/gpuweb#2918, by @zoddicus.
-
Explicitly state the accuracy for ldexp gpuweb/gpuweb#2917, by @zoddicus.
-
Allow control flow reconvergence for short circuit expressions gpuweb/gpuweb#2912, by @alan-baker.
-
Update constructors in GPUError IDL gpuweb/gpuweb#2911, by @toji.
-
Clarify some error types gpuweb/gpuweb#2909, by @alan-baker.
-
Issue a warning on the overheads of zero initializing uninitialized data gpuweb/gpuweb#2907, by @Jiawei-Shao.
-
Use table formatting for texture builtins. gpuweb/gpuweb#2906, by @dj2.
-
Allow reading back from canvases after present gpuweb/gpuweb#2905, by @kainino0x.
-
Forbid spelling non-instantiable pointer types gpuweb/gpuweb#2904, by @dneto0.
-
Clairfy attribute numeric params gpuweb/gpuweb#2902, by @dneto0.
-
Clarify non-normative wording about i32 inference gpuweb/gpuweb#2901, by @dneto0.
-
Clarify fract for small negative numbers gpuweb/gpuweb#2900, by @dneto0.
-
Minor atomic method editorial changes. gpuweb/gpuweb#2898, by @dj2.
-
Add accuracy definition for dot builtin gpuweb/gpuweb#2897, by @zoddicus.
-
Add definitions for bind group descriptors gpuweb/gpuweb#2893, by @toji.
-
Typo L10578 fixed gpuweb/gpuweb#2889, by @abdo643-HULK.
-
Throw when using formats without required feature gpuweb/gpuweb#2887, by @toji.
-
Clarify flat interpoation requirements gpuweb/gpuweb#2885, by @alan-baker.
-
Remove multisampled text from textureNumLayers gpuweb/gpuweb#2882, by @alan-baker.
-
Address out-of-bounds sample index in textureLoad gpuweb/gpuweb#2880, by @alan-baker.
-
Note that face/cull modes don't affect points, lines gpuweb/gpuweb#2878, by @toji.
-
Make GPUShaderModuleCompilationHint.layout optional gpuweb/gpuweb#2876, by @Kangz.
-
Normalizing object creation and invalidation gpuweb/gpuweb#2873, by @toji.
-
wgsl: Make float builtin preconditions similar to integer builtins. gpuweb/gpuweb#2868, by @dj2.
-
editorial: Change 'pointee type' to 'store type' gpuweb/gpuweb#2867, by @dneto0.
-
Formal parameter of ptr type must be instantiable gpuweb/gpuweb#2866, by @dneto0.
-
Relax the null buffer check for zero-attribute vertex buffer slots gpuweb/gpuweb#2865, by @takahirox.
-
Clarify convertability of constant-time expressions gpuweb/gpuweb#2862, by @dneto0.
-
Add algorithm for onSubmittedWorkDone() gpuweb/gpuweb#2861, by @toji.
-
Add table detailing vertex formats gpuweb/gpuweb#2860, by @toji.
-
Even more descriptions for dictionary and enum members gpuweb/gpuweb#2858, by @toji.
-
Add maxInterStageShaderVariables and inter-stage limit validation gpuweb/gpuweb#2857, by @Jiawei-Shao.
-
Support building multi-file specs with the Bikeshed API gpuweb/gpuweb#2856, by @kainino0x.
-
Move "expire stale external textures" slightly earlier gpuweb/gpuweb#2855, by @kainino0x.
-
Clarify references to "update the rendering", "interface object" gpuweb/gpuweb#2854, by @kainino0x.
-
Change GPUError to a base class instead of a union gpuweb/gpuweb#2853, by @kainino0x.
-
Explicitly mark getBindGroupLayout as returning a [NewObject] gpuweb/gpuweb#2852, by @toji.
-
ConversionRank table gets vector, matrix rows gpuweb/gpuweb#2851, by @dneto0.
-
Clarify conversion rank for composites of abstract types gpuweb/gpuweb#2848, by @alan-baker.
-
Validations for writeBuffer and writeTexture should be at device timeline gpuweb/gpuweb#2847, by @Richard-Yunchao.
-
Special case uniformity of arrayLength gpuweb/gpuweb#2846, by @alan-baker.
-
Adopt n11n note from i18n specdev gpuweb/gpuweb#2845, by @mehmetoguzderin.
-
Handle discontinous vertex buffer slots in validation gpuweb/gpuweb#2844, by @jiangzhaoming.
-
Validation for copyExternalImageToTexture should be at queue timeline gpuweb/gpuweb#2842, by @Richard-Yunchao.
-
Explicitly specify precision for color space conversions gpuweb/gpuweb#2841, by @kainino0x.
-
Add guidance on what browsers "should" put in error messages/console gpuweb/gpuweb#2840, by @kainino0x.
-
Add detail to note on viewFormats performance gpuweb/gpuweb#2839, by @kainino0x.
-
Add "bgra8unorm-storage" feature flag gpuweb/gpuweb#2838, by @kainino0x.
-
buffer/device.destroy() unmaps buffer on current thread gpuweb/gpuweb#2837, by @kainino0x.
-
editorial: Define trunc in terms of truncate gpuweb/gpuweb#2836, by @dneto0.
-
Specify computed values for modf gpuweb/gpuweb#2835, by @dneto0.
-
Raise maxComputeWorkgroupStorageSize from 16352 to 16384 gpuweb/gpuweb#2834, by @kainino0x.
-
Reserve binding_array gpuweb/gpuweb#2833, by @dneto0.
-
Add descriptions for buffer and texture creation options gpuweb/gpuweb#2831, by @toji.
-
nit: fix quotation marks in javascript example gpuweb/gpuweb#2827, by @kainino0x.
-
Remove GPUCanvasConfiguration.size gpuweb/gpuweb#2826, by @kainino0x.
-
Add definitions to more render pipeline descriptors gpuweb/gpuweb#2825, by @toji.
-
Remove multithreading support from the spec (for 1.0) gpuweb/gpuweb#2824, by @kainino0x.
-
Validate unknown texture usage bits gpuweb/gpuweb#2823, by @kainino0x.
-
Validate canvas configuration in configure() gpuweb/gpuweb#2821, by @kainino0x.
-
Add a limit for maxColorAttachments gpuweb/gpuweb#2820, by @kainino0x.
-
Factor out "invalidate the current texture", misc cleanups gpuweb/gpuweb#2819, by @kainino0x.
-
Add definitions to all Depth/Stencil State members gpuweb/gpuweb#2817, by @toji.
-
Clarify workgroup variable initialization value description gpuweb/gpuweb#2816, by @alan-baker.
-
Fix second form of mix gpuweb/gpuweb#2814, by @alan-baker.
-
-
89 pull requests merged.
-
Specify the meaning of `Foo.Interface`. gpuweb/gpuweb#2928.
-
AbstractFloat to out-of-finite-range is shader creation error gpuweb/gpuweb#2984.
-
Disallow accessing padding gpuweb/gpuweb#2987.
-
Clarify that evaluations to inf/nan may yield undef value gpuweb/gpuweb#2777.
-
Fix and clean biblio references gpuweb/gpuweb#2952.
-
Describe WGSL's context-aware tokenization gpuweb/gpuweb#2735.
-
Remove commas after final member of a WebIDL enum. gpuweb/gpuweb#2966.
-
Fix proposed IDL in adapters identifier explainer gpuweb/gpuweb#2963.
-
Change ref<S,T,A> to ref<AS,T,AM> gpuweb/gpuweb#2944.
-
'utilize' -> 'use' gpuweb/gpuweb#2954.
-
[wgsl]: Remove interpolation tokens from reserved words gpuweb/gpuweb#2925.
-
Better specify color space conversion elision gpuweb/gpuweb#2956.
-
OOB check in draw calls should ignore empty vertex buffer slots gpuweb/gpuweb#2935.
-
Editorial nits in depth24plus section gpuweb/gpuweb#2949.
-
Allow reading back from canvases after present gpuweb/gpuweb#2905.
-
Clarify 'S' in the ref<S,T,A> row of the ConversionRank gpuweb/gpuweb#2942.
-
[wgsl]: Minor grammar cleanup, add spec link. gpuweb/gpuweb#2937.
-
Fix cross-reference to pointer type gpuweb/gpuweb#2932.
-
Allow control flow reconvergence for short circuit expressions gpuweb/gpuweb#2912.
-
Special case uniformity of arrayLength gpuweb/gpuweb#2846.
-
Make GPUShaderModuleCompilationHint.layout optional gpuweb/gpuweb#2876.
-
Add GPUAdaperInfo to the spec gpuweb/gpuweb#2660.
-
Correct sense of validation test in GPURenderBundleEncoder.finish. gpuweb/gpuweb#2922.
-
Move "expire stale external textures" slightly earlier gpuweb/gpuweb#2855.
-
Issue a warning on the overheads of zero initializing uninitialized data gpuweb/gpuweb#2907.
-
Clarify some error types gpuweb/gpuweb#2909.
-
Update constructors in GPUError IDL gpuweb/gpuweb#2911.
-
Forbid spelling non-instantiable pointer types gpuweb/gpuweb#2904.
-
Clairfy attribute numeric params gpuweb/gpuweb#2902.
-
Clarify non-normative wording about i32 inference gpuweb/gpuweb#2901.
-
Clarify fract for small negative numbers gpuweb/gpuweb#2900.
-
ConversionRank table gets vector, matrix rows gpuweb/gpuweb#2851.
-
Add accuracy definition for dot builtin gpuweb/gpuweb#2897.
-
Minor atomic method editorial changes. gpuweb/gpuweb#2898.
-
Add definitions for bind group descriptors gpuweb/gpuweb#2893.
-
Add table detailing vertex formats gpuweb/gpuweb#2860.
-
Clarity how to interpret GPURenderPassColorAttachment/clearValue gpuweb/gpuweb#2795.
-
Remove GPUCanvasConfiguration.size gpuweb/gpuweb#2826.
-
Validate canvas configuration in configure() gpuweb/gpuweb#2821.
-
Change GPUError to a base class instead of a union gpuweb/gpuweb#2853.
-
Raise maxComputeWorkgroupStorageSize from 16352 to 16384 gpuweb/gpuweb#2834.
-
buffer/device.destroy() unmaps buffer on current thread gpuweb/gpuweb#2837.
-
Move preferred format queries to navigator.gpu gpuweb/gpuweb#2771.
-
Throw when using formats without required feature gpuweb/gpuweb#2887.
-
Note that face/cull modes don't affect points, lines gpuweb/gpuweb#2878.
-
Clarify conversions: floating, and out-of-range AbstractInt gpuweb/gpuweb#2773.
-
Clarify flat interpoation requirements gpuweb/gpuweb#2885.
-
editorial: Change 'pointee type' to 'store type' gpuweb/gpuweb#2867.
-
wgsl: Make float builtin preconditions similar to integer builtins. gpuweb/gpuweb#2868.
-
Clarify convertability of constant-time expressions gpuweb/gpuweb#2862.
-
Remove multisampled text from textureNumLayers gpuweb/gpuweb#2882.
-
Address out-of-bounds sample index in textureLoad gpuweb/gpuweb#2880.
-
Add definitions to more render pipeline descriptors gpuweb/gpuweb#2825.
-
Normalizing object creation and invalidation gpuweb/gpuweb#2873.
-
Add algorithm for onSubmittedWorkDone() gpuweb/gpuweb#2861.
-
Even more descriptions for dictionary and enum members gpuweb/gpuweb#2858.
-
Handle discontinous vertex buffer slots in validation gpuweb/gpuweb#2844.
-
Formal parameter of ptr type must be instantiable gpuweb/gpuweb#2866.
-
Clarify workgroup variable initialization value description gpuweb/gpuweb#2816.
-
Add guidance on what browsers "should" put in error messages/console gpuweb/gpuweb#2840.
-
Adopt n11n note from i18n specdev gpuweb/gpuweb#2845.
-
Clarify references to "update the rendering", "interface object" gpuweb/gpuweb#2854.
-
Explicitly mark getBindGroupLayout as returning a [NewObject] gpuweb/gpuweb#2852.
-
Clarify conversion rank for composites of abstract types gpuweb/gpuweb#2848.
-
Validations for writeBuffer and writeTexture should be at device timeline gpuweb/gpuweb#2847.
-
Specify computed values for modf gpuweb/gpuweb#2835.
-
editorial: Define trunc in terms of truncate gpuweb/gpuweb#2836.
-
Reserve binding_array gpuweb/gpuweb#2833.
-
Explicitly specify precision for color space conversions gpuweb/gpuweb#2841.
-
Add detail to note on viewFormats performance gpuweb/gpuweb#2839.
-
Add "bgra8unorm-storage" feature flag gpuweb/gpuweb#2838.
-
Add validation for GPURender/ComputePassTimestampLocation gpuweb/gpuweb#2812.
-
Add descriptions for buffer and texture creation options gpuweb/gpuweb#2831.
-
Add staticAssert gpuweb/gpuweb#2759.
-
Fix second form of mix gpuweb/gpuweb#2814.
-
Allow creation-time expression in case seletors; Add missings `@const`s gpuweb/gpuweb#2792.
-
nit: fix quotation marks in javascript example gpuweb/gpuweb#2827.
-
Validate unknown texture usage bits gpuweb/gpuweb#2823.
-
Remove multithreading support from the spec (for 1.0) gpuweb/gpuweb#2824.
-
Make use of automatic pipeline layouts explicit gpuweb/gpuweb#2657.
-
Add a limit for maxColorAttachments gpuweb/gpuweb#2820.
-
Factor out "invalidate the current texture", misc cleanups gpuweb/gpuweb#2819.
-
[wgsl] Split regex apart by segment. gpuweb/gpuweb#2762.
-
Specify that filtering with comparison is implementation-defined gpuweb/gpuweb#2425.
-
Add definitions to all Depth/Stencil State members gpuweb/gpuweb#2817.
-
In type rule for boolean negation (`!`), use `T` in precondition. gpuweb/gpuweb#2797.
-
Make IO-shareable types apply to user-defined IO only. gpuweb/gpuweb#2813.
-
Change integer texture parameters to unsigned gpuweb/gpuweb#2722.
-
-
7 pull requests closed.
-
Relax the null buffer check for zero-attribute vertex buffer slots gpuweb/gpuweb#2865.
-
Expose different error types directly instead of via polymorphism gpuweb/gpuweb#2750.
-
Typo L10578 fixed gpuweb/gpuweb#2889.
-
Validation for copyExternalImageToTexture should be at queue timeline gpuweb/gpuweb#2842.
-
Rely on canvas size instead of allowing size to be configured gpuweb/gpuweb#2416.
-
Remove COPY buffer usages gpuweb/gpuweb#2172.
-
Move all canvas configure validation to getCurrentTexture gpuweb/gpuweb#2536.
-
-
78 pull requests received 242 new comments.
-
Add definitions to more render pipeline descriptors gpuweb/gpuweb#2825 (13 comments by @kainino0x, @toji)
-
Add maxInterStageShaderVariables and inter-stage limit validation gpuweb/gpuweb#2857 (11 comments by @Jiawei-Shao, @kainino0x, @Kangz)
-
Clarity how to interpret GPURenderPassColorAttachment/clearValue gpuweb/gpuweb#2795 (10 comments by @Jiawei-Shao, @kainino0x)
-
Add descriptions for buffer and texture creation options gpuweb/gpuweb#2831 (8 comments by @kainino0x, @toji)
-
Use PredefinedColorSpace instead of GPUPredefinedColorSpace gpuweb/gpuweb#2967 (5 comments by @kainino0x, @Kangz)
-
Issue a warning on the overheads of zero initializing uninitialized data gpuweb/gpuweb#2907 (5 comments by @Jiawei-Shao, @kainino0x, @toji)
-
Throw when using formats without required feature gpuweb/gpuweb#2887 (5 comments by @kainino0x, @toji)
-
Add guidance on what browsers "should" put in error messages/console gpuweb/gpuweb#2840 (5 comments by @kainino0x, @toji)
-
AbstractFloat to out-of-finite-range is shader creation error gpuweb/gpuweb#2984 (4 comments by @dneto0, @kdashg)
-
Better specify color space conversion elision gpuweb/gpuweb#2956 (4 comments by @kainino0x, @tabatkins, @toji)
-
Make use of automatic pipeline layouts explicit gpuweb/gpuweb#2657 (4 comments by @juj, @kainino0x, @Kangz, @toji)
-
Even more descriptions for dictionary and enum members gpuweb/gpuweb#2858 (4 comments by @kainino0x)
-
Formal parameter of ptr type must be instantiable gpuweb/gpuweb#2866 (4 comments by @alan-baker, @dneto0, @jimblandy)
-
Add shift operators, remove shiftLeft, shiftRight builtins gpuweb/gpuweb#2989 (3 comments by @dneto0)
-
OOB check in draw calls should ignore empty vertex buffer slots gpuweb/gpuweb#2935 (3 comments by @kainino0x)
-
Change GPUError to a base class instead of a union gpuweb/gpuweb#2853 (3 comments by @kainino0x, @toji)
-
Expose different error types directly instead of via polymorphism gpuweb/gpuweb#2750 (3 comments by @kainino0x, @toji)
-
Clarify conversion rank for composites of abstract types gpuweb/gpuweb#2848 (3 comments by @ben-clayton, @dneto0)
-
Clarify that evaluations to inf/nan may yield undef value gpuweb/gpuweb#2777 (2 comments by @kdashg)
-
ConversionRank table gets vector, matrix rows gpuweb/gpuweb#2851 (2 comments by @alan-baker, @dneto0)
-
editorial: Change 'pointee type' to 'store type' gpuweb/gpuweb#2867 (2 comments by @alan-baker, @dneto0)
-
wgsl: Make float builtin preconditions similar to integer builtins. gpuweb/gpuweb#2868 (2 comments by @alan-baker, @dj2)
-
buffer/device.destroy() unmaps buffer on current thread gpuweb/gpuweb#2837 (2 comments by @kainino0x)
-
Clarify conversions: floating, and out-of-range AbstractInt gpuweb/gpuweb#2773 (2 comments by @dneto0, @kdashg)
-
Clarify workgroup variable initialization value description gpuweb/gpuweb#2816 (2 comments by @kdashg)
-
Clarify convertability of constant-time expressions gpuweb/gpuweb#2862 (2 comments by @alan-baker, @dneto0)
-
Add validation for GPURender/ComputePassTimestampLocation gpuweb/gpuweb#2812 (2 comments by @haoxli, @kainino0x)
-
Allow control flow reconvergence for short circuit expressions gpuweb/gpuweb#2912 (1 comment by @kdashg)
-
Relax the null buffer check for zero-attribute vertex buffer slots gpuweb/gpuweb#2865 (1 comment by @takahirox)
-
[editorial] Change vast majority of must language to link to shader-creation error gpuweb/gpuweb#2746 (1 comment by @alan-baker)
-
Validation for copyExternalImageToTexture should be at queue timeline gpuweb/gpuweb#2842 (1 comment by @kainino0x)
-
Allow creation-time expression in case seletors; Add missings `@const`s gpuweb/gpuweb#2792 (1 comment by @dneto0)
-
Rely on canvas size instead of allowing size to be configured gpuweb/gpuweb#2416 (1 comment by @kainino0x)
-
Move all canvas configure validation to getCurrentTexture gpuweb/gpuweb#2536 (1 comment by @kainino0x)
-
Specify that filtering with comparison is implementation-defined gpuweb/gpuweb#2425 (1 comment by @kainino0x)
-
In type rule for boolean negation (`!`), use `T` in precondition. gpuweb/gpuweb#2797 (1 comment by @jimblandy)
-
gpuweb/cts
-
71 issues opened.
-
[wgsl] padding bytes must not be accessed gpuweb/cts#1498, by @dneto0.
-
wgsl: validate `vec` parsing gpuweb/cts#1495, by @dj2.
-
wgsl: validate `mat` parsing gpuweb/cts#1494, by @dj2.
-
wgsl: validate `atomic` parsing gpuweb/cts#1493, by @dj2.
-
wgsl: Validate `array` parsing gpuweb/cts#1492, by @dj2.
-
wgsl: Validate array type comparisions gpuweb/cts#1491, by @dj2.
-
wgsl: validate array creation gpuweb/cts#1490, by @dj2.
-
wgsl: execution tests for ordering of atomic writes gpuweb/cts#1489, by @dj2.
-
wgsl: validate all arithmetic operations fail if one of the types is an atomic. gpuweb/cts#1488, by @dj2.
-
wgsl: Execution tests for atomic initialization gpuweb/cts#1487, by @dj2.
-
wgsl: validate atomics disallow initializer expressions gpuweb/cts#1486, by @dj2.
-
wgsl: Validate atomic type creation gpuweb/cts#1485, by @dj2.
-
wgsl: Validate matrix creation with non-floating point types gpuweb/cts#1484, by @dj2.
-
wgsl: Validate vector creation with non scalar and non-abstract types fails gpuweb/cts#1483, by @dj2.
-
wgsl: Testing for large abstract values gpuweb/cts#1482, by @dj2.
-
Provide ability to dump WGSL shaders gpuweb/cts#1479, by @dj2.
-
wgsl: validate conversion rules gpuweb/cts#1478, by @dj2.
-
Organize all tests? gpuweb/cts#1477, by @kainino0x.
-
web_platform: Add GPUExternalTexuture.expired test cases gpuweb/cts#1473, by @shaoboyan.
-
wgsl: test hex float literal exact representation as f32 gpuweb/cts#1472, by @dneto0.
-
wgsl: Test small magnitude hex floats. gpuweb/cts#1470, by @dj2.
-
wgsl: Test small magnitude floats gpuweb/cts#1469, by @dj2.
-
wgsl: execution tests for scoping rules gpuweb/cts#1468, by @dj2.
-
wgsl: Validate recursion disallowed gpuweb/cts#1467, by @dj2.
-
wgsl: validate scoping rules gpuweb/cts#1466, by @dj2.
-
wgsl: validate `global_directive` gpuweb/cts#1465, by @dj2.
-
wgsl: validate pipeline stage attributes gpuweb/cts#1464, by @dj2.
-
wgsl: execution tests for `@workgroup_size` gpuweb/cts#1463, by @dj2.
-
wgsl: validate `@workgroup_size` parsing gpuweb/cts#1462, by @dj2.
-
wgsl: Add shader regression tests for countOneBits and reverseBits for HLSL gpuweb/cts#1458, by @amaiorano.
-
wgsl: validate all examples in the spec exist in the CTS gpuweb/cts#1457, by @dj2.
-
wgsl: Validate various structs with size/align validate correctly. gpuweb/cts#1456, by @dj2.
-
wgsl: validate `@size` parsing gpuweb/cts#1455, by @dj2.
-
wgsl: execution tests for `@location` gpuweb/cts#1454, by @dj2.
-
wgsl: validate `@location` parsing gpuweb/cts#1453, by @dj2.
-
wgsl: execution test for `@invariant` gpuweb/cts#1452, by @dj2.
-
wgsl: execution tests for `@interpolate` gpuweb/cts#1451, by @dj2.
-
wgsl: Validate parsing of the `@interpolate` builtin gpuweb/cts#1449, by @dj2.
-
wgsl: validate `@invariant` parsing gpuweb/cts#1447, by @dj2.
-
wgsl: validate `@builtin` parsing gpuweb/cts#1446, by @dj2.
-
wgsl: Execution tests for `@id` gpuweb/cts#1445, by @dj2.
-
wgsl: Validate `@id` parsing gpuweb/cts#1444, by @dj2.
-
wgsl: Validate each `@builtin` value can be used as variable names gpuweb/cts#1443, by @dj2.
-
wgsl: updated error conditions for vector accessing gpuweb/cts#1441, by @dneto0.
-
wgsl: Add execution tests for `@binding` and `@group` values gpuweb/cts#1440, by @dj2.
-
wgsl: Validate `@group` parsing gpuweb/cts#1439, by @dj2.
-
wgsl: Validate `@const` is not permitted gpuweb/cts#1438, by @dj2.
-
wgsl: Add builtin validation test for `frag_depth` gpuweb/cts#1437, by @dj2.
-
wgsl: Add `@builtin` execution tests gpuweb/cts#1436, by @dj2.
-
wgsl: validate `@binding` parsing gpuweb/cts#1434, by @dj2.
-
wgsl: `@align` execution tests gpuweb/cts#1433, by @dj2.
-
wgsl: Validate `@align` parsing gpuweb/cts#1432, by @dj2.
-
wgsl: Validate a line ending comment can terminate at end of program. gpuweb/cts#1431, by @dj2.
-
wgsl: Test line break characters gpuweb/cts#1430, by @dj2.
-
wgsl: Non UTF-8 text is a valid encoding. gpuweb/cts#1429, by @dj2.
-
wgsl: Shader does not require an entry point gpuweb/cts#1428, by @dj2.
-
forbid spelling pointer types that are not instantiable gpuweb/cts#1427, by @dneto0.
-
test case: invalid to have bare type constructor or type conversion gpuweb/cts#1411, by @dneto0.
-
`insertLeaf` not setting `key` correctly. gpuweb/cts#1409, by @dj2.
-
web_platform: import from 100 different video streams simultaneously gpuweb/cts#1398, by @kainino0x.
-
wgsl: Formal parameter of pointer type must be instantiable as a variable gpuweb/cts#1395, by @dneto0.
-
wgsl: `anyOf(...)` should support `Comparator`s gpuweb/cts#1391, by @zoddicus.
-
wgsl: execution test for textureGather builtin gpuweb/cts#1377, by @dj2.
-
wgsl: Make handling inputs potentially being flushed easier when writing tests gpuweb/cts#1341, by @zoddicus.
-
wgsl: `atan2` tests do not account for backends performing FTZ on inputs gpuweb/cts#1339, by @zoddicus.
-
-
11 issues closed.
-
wgsl: Test small magnitude hex floats. gpuweb/cts#1470.
-
Update tests for sparse render attachments if added gpuweb/cts#901.
-
`insertLeaf` not setting `key` correctly. gpuweb/cts#1409.
-
Add comprehensive memory model tests gpuweb/cts#963.
-
wgsl: `anyOf(...)` should support `Comparator`s gpuweb/cts#1391.
-
correctlyRoundedThreshold does not correctly handling some cases gpuweb/cts#1332.
-
27 issues received 40 new comments.
-
wgsl: test hex float literal exact representation as f32 gpuweb/cts#1472 (5 comments by @dj2, @dneto0)
-
web_platform: copyExternalImageToTexture with more formats, extended-srgb and other color spaces gpuweb/cts#913 (2 comments by @kainino0x, @shaoboyan)
-
wgsl: validate you can declare things with names that are: attributes, named params for attrib, texel format names gpuweb/cts#1448 (2 comments by @dj2, @dneto0)
-
wgsl: Validate vector creation with non scalar and non-abstract types fails gpuweb/cts#1483 (1 comment by @dj2)
-
wgsl: Add shader regression tests for countOneBits and reverseBits for HLSL gpuweb/cts#1458 (1 comment by @amaiorano)
-
wgsl: Validate `@builtin` on non-entrypoint parameter, non-entrypoint return type or struct member fails gpuweb/cts#1435 (1 comment by @dj2)
-
wgsl: Validate a line ending comment can terminate at end of program. gpuweb/cts#1431 (1 comment by @dj2)
-
web_platform: import from 100 different video streams simultaneously gpuweb/cts#1398 (1 comment by @kainino0x)
-
-
91 pull requests opened.
-
Factor copyToTexture:canvas tests common code. gpuweb/cts#1497, by @Kangz.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part IV gpuweb/cts#1496, by @Jiawei-Shao.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part III gpuweb/cts#1480, by @Jiawei-Shao.
-
Draw buffer validation test for unused and overlapped cases gpuweb/cts#1475, by @jiangzhaoming.
-
Change importExternalTexture,destroy to importExternalTexture,expired gpuweb/cts#1474, by @shaoboyan.
-
Revise ExternalTexture Video Test Expectations gpuweb/cts#1471, by @bjjones.
-
Draw vertex buffer validation for instance step mode buffer gpuweb/cts#1461, by @jiangzhaoming.
-
resource_init: Specify view dimension in checkContentsBySampling gpuweb/cts#1460, by @Kangz.
-
Fix memory leak in vertex_buffer_OOB,vertex_step_mode gpuweb/cts#1459, by @jiangzhaoming.
-
Draw vertex buffer access test for vertex step mode buffer gpuweb/cts#1450, by @jiangzhaoming.
-
wgsl: Stub synchronization builtin functions. gpuweb/cts#1426, by @dj2.
-
wgsl: Add stubs for the pack and unpack methods. gpuweb/cts#1425, by @dj2.
-
Duplicate params: catch in ParamsBuilder, improve error in loadTreeForQuery gpuweb/cts#1424, by @kainino0x.
-
wgsl: Stub tests for the textureStore builtin. gpuweb/cts#1423, by @dj2.
-
wgsl: Add stub tests for textureSampleLevel. gpuweb/cts#1422, by @dj2.
-
wgsl: Add stub tests for textureSampleGrad. gpuweb/cts#1421, by @dj2.
-
wgsl: Add stub tests for textureSampleCompareLevel. gpuweb/cts#1420, by @dj2.
-
wgsl: Stub tests for textureSampleCompare gpuweb/cts#1418, by @dj2.
-
wgsl: Add stubs for the read-write-modify atomic operations. gpuweb/cts#1417, by @dj2.
-
wgsl: Add stubs for the atomicStore builtin. gpuweb/cts#1416, by @dj2.
-
wgsl: Stub tests for the atomicLoad builtin. gpuweb/cts#1415, by @dj2.
-
wgsl: Add stub tests for the textureSample builtin. gpuweb/cts#1414, by @dj2.
-
Add skipLibCheck to tsconfig.json gpuweb/cts#1413, by @dj2.
-
Add all "recommended" typescript-eslint lints gpuweb/cts#1412, by @kainino0x.
-
wgsl: Fixup coord parameters in texture tests. gpuweb/cts#1410, by @dj2.
-
Fix tests using DS attachments not encompassing all aspects. gpuweb/cts#1408, by @Kangz.
-
attachment_compatibility: Properly set d/sReadOnly when testing it. gpuweb/cts#1407, by @Kangz.
-
wgsl: Add stub tests for textureNumSamples gpuweb/cts#1406, by @dj2.
-
wgsl: Stub tests for textureNumLevels builtin. gpuweb/cts#1405, by @dj2.
-
wgsl: Stub tests for textureNumLayers builtin gpuweb/cts#1404, by @dj2.
-
wgsl: Stub tests for textureLoad builtin. gpuweb/cts#1403, by @dj2.
-
DNS: Prototype for how to implement inherited FP accuracy tests gpuweb/cts#1402, by @zoddicus.
-
Remove dupes in anyOf to clean up error messages gpuweb/cts#1401, by @amaiorano.
-
Fix WGSL deprecation warnings in operation/rendering/draw.spec.ts gpuweb/cts#1400, by @Kangz.
-
Replace flushSubnormalNumber with isSubnormalNumber gpuweb/cts#1399, by @amaiorano.
-
Fix unary tests with subnormal inputs gpuweb/cts#1397, by @amaiorano.
-
wgsl: Cleanup duplicate description text gpuweb/cts#1396, by @dj2.
-
wgsl: Use `kValue.`... instead of converting bit-fields to `Scalar`s gpuweb/cts#1394, by @zoddicus.
-
wgsl: Extend `anyOf` to accept `Comparator`s gpuweb/cts#1393, by @zoddicus.
-
wgsl: Stub textureSampleBias tests gpuweb/cts#1392, by @dj2.
-
Raise maxComputeWorkgroupStorageSize from 16352 to 16384 gpuweb/cts#1390, by @austinEng.
-
wgsl: update textureGather tests gpuweb/cts#1389, by @dj2.
-
Output whether value is subnormal in Scalar.toString() gpuweb/cts#1388, by @amaiorano.
-
Fix atan2 test failures when Y is subnormal gpuweb/cts#1387, by @amaiorano.
-
Reland "Simplify device selection" with fix gpuweb/cts#1386, by @kainino0x.
-
Add handling of non-unique results in `atan2` tests gpuweb/cts#1385, by @zoddicus.
-
Revert "Simplify device selection" and dependent commits gpuweb/cts#1384, by @kainino0x.
-
nit: remove dead code gpuweb/cts#1383, by @kainino0x.
-
[dawn-node]: Move TSC config to file. gpuweb/cts#1382, by @dj2.
-
Fix crash in GPUTest.init awaiting provider gpuweb/cts#1381, by @kainino0x.
-
Fix order in package.json gpuweb/cts#1380, by @jiangzhaoming.
-
wgsl: Adds stub tests for textureGatherCompare gpuweb/cts#1379, by @dj2.
-
wgsl: Stub tests for the textureGather builtin. gpuweb/cts#1378, by @dj2.
-
Fix skipping tests when device isn't available gpuweb/cts#1376, by @kainino0x.
-
wgsl: Stub tests for textureDimension. gpuweb/cts#1375, by @dj2.
-
[wgsl] Add missing @const decorations. gpuweb/cts#1374, by @dj2.
-
[wgs] fixup mix signature to match spec. gpuweb/cts#1373, by @dj2.
-
[wgs] sync modf text with spec update. gpuweb/cts#1372, by @dj2.
-
[wgsl] reserve binding_array. gpuweb/cts#1371, by @dj2.
-
Explicitly specify 2d texture views in some cases gpuweb/cts#1370, by @toji.
-
wgsl: Add support for f64 to Scalar gpuweb/cts#1369, by @zoddicus.
-
Allow eventualExpectOk to accept Promise.all gpuweb/cts#1368, by @shrekshao.
-
wgsl: Generalize unary and binary f32 arithmetic case generation gpuweb/cts#1367, by @zoddicus.
-
wgsl: Add missing parameterizations gpuweb/cts#1366, by @dj2.
-
wgsl: Stub tests for the fwidthFine builtin. gpuweb/cts#1365, by @dj2.
-
wgsl: Stub tests for the fwidthCoarse builtin. gpuweb/cts#1364, by @dj2.
-
wgsl: Stub tests for the fwidth builtin. gpuweb/cts#1363, by @dj2.
-
wgsl: Stub tests for the dpdyFine builtin. gpuweb/cts#1362, by @dj2.
-
wgsl: Stub tests for the dpdyCoarse builtin. gpuweb/cts#1361, by @dj2.
-
wgsl: Stub tests for the dpdy builtin gpuweb/cts#1360, by @dj2.
-
wgsl: Stub tests for the dpdxFine builtin. gpuweb/cts#1359, by @dj2.
-
wgsl: Stub tests for dpdxCoarse builtin gpuweb/cts#1358, by @dj2.
-
wgsl: Stub tests for dpdx builtin gpuweb/cts#1357, by @dj2.
-
wgsl: Stub tests for the dot builtin. gpuweb/cts#1356, by @dj2.
-
wgsl: Stub tests for the transpose builtin. gpuweb/cts#1354, by @dj2.
-
wgsl: Stub tests for determinant builtin. gpuweb/cts#1353, by @dj2.
-
wgsl: Stub tests for trunc builtin. gpuweb/cts#1352, by @dj2.
-
wgsl: Stub tests for the tanh builtin. gpuweb/cts#1351, by @dj2.
-
wgsl: Stub tests for tan builtin. gpuweb/cts#1350, by @dj2.
-
wgsl: Stub tests for step builtin. gpuweb/cts#1349, by @dj2.
-
wgsl: Stub tests for sqrt builtin. gpuweb/cts#1348, by @dj2.
-
wgsl: Stub tests for smoothstep builtin. gpuweb/cts#1347, by @dj2.
-
wgsl: Stub tests for sinh builtin. gpuweb/cts#1346, by @dj2.
-
wgsl: Stub for sign builtin. gpuweb/cts#1345, by @dj2.
-
Sparse color attachment fix gpuweb/cts#1344, by @shrekshao.
-
DevicePool: merge default holder into the DescriptorToHolderMap gpuweb/cts#1343, by @kainino0x.
-
remove doInit/doFinalize gpuweb/cts#1342, by @kainino0x.
-
Handle backends flushing inputs in `atan2` tests gpuweb/cts#1340, by @zoddicus.
-
Add --colors flag for enabling color output gpuweb/cts#1338, by @ben-clayton.
-
Device selection cleanup, etc gpuweb/cts#1337, by @kainino0x.
-
Fix non-object-type test-parameter values gpuweb/cts#1336, by @kainino0x.
-
-
90 pull requests merged.
-
Draw buffer validation test for unused and overlapped cases gpuweb/cts#1475.
-
Revise ExternalTexture Video Test Expectations gpuweb/cts#1471.
-
Draw vertex buffer validation for instance step mode buffer gpuweb/cts#1461.
-
resource_init: Specify view dimension in checkContentsBySampling gpuweb/cts#1460.
-
Fix memory leak in vertex_buffer_OOB,vertex_step_mode gpuweb/cts#1459.
-
Draw vertex buffer access test for vertex step mode buffer gpuweb/cts#1450.
-
Add all "recommended" typescript-eslint lints gpuweb/cts#1412.
-
Raise maxComputeWorkgroupStorageSize from 16352 to 16384 gpuweb/cts#1390.
-
wgsl: Add stubs for the pack and unpack methods. gpuweb/cts#1425.
-
wgsl: Stub tests for the textureStore builtin. gpuweb/cts#1423.
-
wgsl: Add stub tests for textureSampleLevel. gpuweb/cts#1422.
-
wgsl: Stub synchronization builtin functions. gpuweb/cts#1426.
-
Update pipeline output targets validation with different writeMask gpuweb/cts#1174.
-
wgsl: Add stub tests for textureSampleGrad. gpuweb/cts#1421.
-
wgsl: Add stub tests for textureSampleCompareLevel. gpuweb/cts#1420.
-
wgsl: Stub tests for textureSampleCompare gpuweb/cts#1418.
-
wgsl: Add stubs for the read-write-modify atomic operations. gpuweb/cts#1417.
-
wgsl: Add stubs for the atomicStore builtin. gpuweb/cts#1416.
-
wgsl: Stub tests for the atomicLoad builtin. gpuweb/cts#1415.
-
wgsl: Add stub tests for the textureSample builtin. gpuweb/cts#1414.
-
wgsl: Fixup coord parameters in texture tests. gpuweb/cts#1410.
-
Add skipLibCheck to tsconfig.json gpuweb/cts#1413.
-
attachment_compatibility: Properly set d/sReadOnly when testing it. gpuweb/cts#1407.
-
Fix tests using DS attachments not encompassing all aspects. gpuweb/cts#1408.
-
Remove dupes in anyOf to clean up error messages gpuweb/cts#1401.
-
Fix WGSL deprecation warnings in operation/rendering/draw.spec.ts gpuweb/cts#1400.
-
wgsl: Add stub tests for textureNumSamples gpuweb/cts#1406.
-
wgsl: Stub tests for textureNumLevels builtin. gpuweb/cts#1405.
-
wgsl: Stub tests for textureNumLayers builtin gpuweb/cts#1404.
-
wgsl: Stub tests for textureLoad builtin. gpuweb/cts#1403.
-
Replace flushSubnormalNumber with isSubnormalNumber gpuweb/cts#1399.
-
Fix unary tests with subnormal inputs gpuweb/cts#1397.
-
[dawn-node]: Move TSC config to file. gpuweb/cts#1382.
-
wgsl: Cleanup duplicate description text gpuweb/cts#1396.
-
Add comprehensive memory model tests gpuweb/cts#1330.
-
Reland "Simplify device selection" with fix gpuweb/cts#1386.
-
Output whether value is subnormal in Scalar.toString() gpuweb/cts#1388.
-
wgsl: Stub textureSampleBias tests gpuweb/cts#1392.
-
wgsl: Use `kValue.`... instead of converting bit-fields to `Scalar`s gpuweb/cts#1394.
-
wgsl: update textureGather tests gpuweb/cts#1389.
-
wgsl: Extend `anyOf` to accept `Comparator`s gpuweb/cts#1393.
-
Fix atan2 test failures when Y is subnormal gpuweb/cts#1387.
-
Revert "Simplify device selection" and dependent commits gpuweb/cts#1384.
-
nit: remove dead code gpuweb/cts#1383.
-
wgsl: Add support for f64 to Scalar gpuweb/cts#1369.
-
Fix crash in GPUTest.init awaiting provider gpuweb/cts#1381.
-
Fix order in package.json gpuweb/cts#1380.
-
wgsl: Stub tests for the textureGather builtin. gpuweb/cts#1378.
-
wgsl: Stub tests for textureDimension. gpuweb/cts#1375.
-
wgsl: Adds stub tests for textureGatherCompare gpuweb/cts#1379.
-
Fix skipping tests when device isn't available gpuweb/cts#1376.
-
[wgsl] Add missing @const decorations. gpuweb/cts#1374.
-
[wgs] fixup mix signature to match spec. gpuweb/cts#1373.
-
[wgs] sync modf text with spec update. gpuweb/cts#1372.
-
[wgsl] reserve binding_array. gpuweb/cts#1371.
-
Explicitly specify 2d texture views in some cases gpuweb/cts#1370.
-
Device selection cleanup, etc gpuweb/cts#1337.
-
Allow eventualExpectOk to accept Promise.all gpuweb/cts#1368.
-
wgsl: Generalize unary and binary f32 arithmetic case generation gpuweb/cts#1367.
-
wgsl: Add missing parameterizations gpuweb/cts#1366.
-
wgsl: Stub tests for the fwidthFine builtin. gpuweb/cts#1365.
-
wgsl: Stub tests for the fwidthCoarse builtin. gpuweb/cts#1364.
-
wgsl: Stub tests for the fwidth builtin. gpuweb/cts#1363.
-
wgsl: Stub tests for the dpdyFine builtin. gpuweb/cts#1362.
-
wgsl: Stub tests for the dpdyCoarse builtin. gpuweb/cts#1361.
-
wgsl: Stub tests for the dpdy builtin gpuweb/cts#1360.
-
wgsl: Stub tests for the dpdxFine builtin. gpuweb/cts#1359.
-
wgsl: Stub tests for dpdxCoarse builtin gpuweb/cts#1358.
-
wgsl: Stub tests for dpdx builtin gpuweb/cts#1357.
-
wgsl: Stub tests for the dot builtin. gpuweb/cts#1356.
-
wgsl: Stub tests for the transpose builtin. gpuweb/cts#1354.
-
wgsl: Stub tests for determinant builtin. gpuweb/cts#1353.
-
wgsl: Stub tests for trunc builtin. gpuweb/cts#1352.
-
wgsl: Stub tests for the tanh builtin. gpuweb/cts#1351.
-
wgsl: Stub tests for tan builtin. gpuweb/cts#1350.
-
wgsl: Stub tests for step builtin. gpuweb/cts#1349.
-
wgsl: Stub tests for sqrt builtin. gpuweb/cts#1348.
-
wgsl: Stub tests for smoothstep builtin. gpuweb/cts#1347.
-
wgsl: Stub tests for sinh builtin. gpuweb/cts#1346.
-
wgsl: Stub for sign builtin. gpuweb/cts#1345.
-
Don't block on subcases until the end of a testcase gpuweb/cts#1335.
-
Sparse color attachment fix gpuweb/cts#1344.
-
DevicePool: merge default holder into the DescriptorToHolderMap gpuweb/cts#1343.
-
remove doInit/doFinalize gpuweb/cts#1342.
-
Handle backends flushing inputs in `atan2` tests gpuweb/cts#1340.
-
Move test device selection to an optional before clause gpuweb/cts#1333.
-
Fix non-object-type test-parameter values gpuweb/cts#1336.
-
Add --colors flag for enabling color output gpuweb/cts#1338.
-
3 pull requests closed.
-
Add an option to run subcases in parallel gpuweb/cts#1169.
-
Add handling of non-unique results in `atan2` tests gpuweb/cts#1385.
-
Pipeline gathering subcase results gpuweb/cts#1172.
-
-
28 pull requests received 174 new comments.
-
DNS: Prototype for how to implement inherited FP accuracy tests gpuweb/cts#1402 (53 comments by @ben-clayton, @dneto0, @zoddicus)
-
Move test device selection to an optional before clause gpuweb/cts#1333 (22 comments by @austinEng, @kainino0x)
-
Revise ExternalTexture Video Test Expectations gpuweb/cts#1471 (19 comments by @bjjones, @kainino0x, @shaoboyan)
-
Remove dupes in anyOf to clean up error messages gpuweb/cts#1401 (7 comments by @amaiorano, @zoddicus)
-
Update pipeline output targets validation with different writeMask gpuweb/cts#1174 (6 comments by @austinEng, @shrekshao)
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part III gpuweb/cts#1480 (4 comments by @jiangzhaoming, @Jiawei-Shao)
-
Draw vertex buffer access test for vertex step mode buffer gpuweb/cts#1450 (4 comments by @jiangzhaoming, @Jiawei-Shao)
-
Fix unary tests with subnormal inputs gpuweb/cts#1397 (4 comments by @amaiorano, @ben-clayton, @zoddicus)
-
Don't block on subcases until the end of a testcase gpuweb/cts#1335 (4 comments by @austinEng, @kainino0x)
-
Duplicate params: catch in ParamsBuilder, improve error in loadTreeForQuery gpuweb/cts#1424 (3 comments by @dj2, @kainino0x)
-
attachment_compatibility: Properly set d/sReadOnly when testing it. gpuweb/cts#1407 (3 comments by @Kangz, @shrekshao)
-
Draw buffer validation test for unused and overlapped cases gpuweb/cts#1475 (2 comments by @jiangzhaoming, @Jiawei-Shao)
-
Draw vertex buffer validation for instance step mode buffer gpuweb/cts#1461 (2 comments by @jiangzhaoming)
-
Change importExternalTexture,destroy to importExternalTexture,expired gpuweb/cts#1474 (1 comment by @shaoboyan)
-
TR publications
-
26 documents published.
-
WebGPU published 13 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 13 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/gpuweb
-
41 issues opened.
-
Clarify atomics initialization for variables in workgroup address space gpuweb/gpuweb#2811, by @teoxoy.
-
Escape Mechanisms in Identifiers gpuweb/gpuweb#2810, by @mehmetoguzderin.
-
Attribute 'Valid Values' should be updated for abstract integers gpuweb/gpuweb#2809, by @ben-clayton.
-
Accuracy of dot built-in never specified gpuweb/gpuweb#2808, by @zoddicus.
-
Return type of second form of `mix` seems incorrect. gpuweb/gpuweb#2807, by @dj2.
-
Third param of `mix` with auto-splat param should be `T`. gpuweb/gpuweb#2806, by @dj2.
-
Double precision floats (IEEE-754 binary64) gpuweb/gpuweb#2805, by @SamueleDecarli.
-
Spec clarification around untyped LHS and abstract RHS gpuweb/gpuweb#2802, by @dj2.
-
`GPUPipelineBase.getBindGroupLayout` definition doesn't cover out-of-range index gpuweb/gpuweb#2801, by @jimblandy.
-
Synchronization Built-in Functions Examples and Refinement gpuweb/gpuweb#2800, by @elyshaffir.
-
Clarify behavior of float division/remainder by 0 gpuweb/gpuweb#2798, by @teoxoy.
-
Consider making `binding_array` a reserved word gpuweb/gpuweb#2794, by @jimblandy.
-
wgsl: extractBits out of alphabetical order in spec text gpuweb/gpuweb#2793, by @dj2.
-
Consider allowing creation-time expressions for attributes gpuweb/gpuweb#2790, by @teoxoy.
-
Add back constraint that DS views must have aspect "all" for rendering. gpuweb/gpuweb#2789, by @Kangz.
-
Missing Abstract Int/Float preconditions for Conversion Expressions gpuweb/gpuweb#2788, by @teoxoy.
-
`@const` missing on some built-in functions gpuweb/gpuweb#2787, by @teoxoy.
-
GPUExternalTexture [[active_imports]] only works if there are RAFs gpuweb/gpuweb#2785, by @Kangz.
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780, by @mehmetoguzderin.
-
Collaborate with the immersive web groups for WebXR interop gpuweb/gpuweb#2778, by @Kangz.
-
Potential parsing conflict with const literals and postfix expressions. gpuweb/gpuweb#2772, by @dj2.
-
Number literal conversion clarification gpuweb/gpuweb#2769, by @teoxoy.
-
New reserved word list includes too many common words gpuweb/gpuweb#2768, by @jimblandy.
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766, by @Kangz.
-
Clarify undefined behavior of built in functions gpuweb/gpuweb#2765, by @teoxoy.
-
Consider allowing creation-time expressions or constants in switch case selectors gpuweb/gpuweb#2764, by @teoxoy.
-
Explicit flat interpolation requirement clarification gpuweb/gpuweb#2760, by @teoxoy.
-
adapt line-ending comment to Unicode conventions gpuweb/gpuweb#2758, by @dneto0.
-
Are f16 hex float literals allowed? gpuweb/gpuweb#2756, by @dj2.
-
Combined limit for max vertex stage bind groups + vertex buffer bindings gpuweb/gpuweb#2749, by @litherum.
-
The preferred format is not always storage-capable gpuweb/gpuweb#2748, by @kainino0x.
-
createRenderPipeline with preferred format is dependent on canvas gpuweb/gpuweb#2747, by @kainino0x.
-
Consider using a more intuitive matrix notation gpuweb/gpuweb#2744, by @teoxoy.
-
WebGPURenderingContext should support preserveDrawingBuffer: true gpuweb/gpuweb#2743, by @greggman.
-
Pointers in uniformity analysis gpuweb/gpuweb#2741, by @alan-baker.
-
dawn_wire crashes on normal ARM devices with SIGBUS (memory alignment issues) gpuweb/gpuweb#2739, by @hefeng-oppo.
-
Missing depth24unorm format? gpuweb/gpuweb#2732, by @Kangz.
-
Missing specification of error type produced gpuweb/gpuweb#2730, by @dj2.
-
Multisample resolve when base texture formats do not match gpuweb/gpuweb#2729, by @austinEng.
-
-
40 issues closed.
-
Allow copying from any canvas that toDataURL can be called on gpuweb/gpuweb#2373.
-
`GPUPipelineBase.getBindGroupLayout` definition doesn't cover out-of-range index gpuweb/gpuweb#2801.
-
New reserved word list includes too many common words gpuweb/gpuweb#2768.
-
Missing Abstract Int/Float preconditions for Conversion Expressions gpuweb/gpuweb#2788.
-
Add back constraint that DS views must have aspect "all" for rendering. gpuweb/gpuweb#2789.
-
MapAsync with subranges. gpuweb/gpuweb#605.
-
Exposing GPU hardware information to web developers in WebGPU gpuweb/gpuweb#2195.
-
GPUAdapter.name considered harmful gpuweb/gpuweb#2191.
-
Should WGSL have different types of extensions? gpuweb/gpuweb#2310.
-
Dealing with CRLF and error line numbers gpuweb/gpuweb#2435.
-
Proposal of FP16 WGSL extension gpuweb/gpuweb#2512.
-
FMod vs FRem semantics gpuweb/gpuweb#1696.
-
call_statement: should only be able to call a function that yields void? gpuweb/gpuweb#1413.
-
Supporting MSAA without rendering capability? gpuweb/gpuweb#2465.
-
Clarify the vertex stage output / fragment stage input matching rules gpuweb/gpuweb#1953.
-
adapt line-ending comment to Unicode conventions gpuweb/gpuweb#2758.
-
FP16 support gpuweb/gpuweb#658.
-
Concerns around error label and message gpuweb/gpuweb#2633.
-
Investigation: Sampling Depth Textures gpuweb/gpuweb#659.
-
my2DArrayTexture.createView() doesn't create a full view gpuweb/gpuweb#2684.
-
Are f16 hex float literals allowed? gpuweb/gpuweb#2756.
-
wgsl: don't reserve certain words gpuweb/gpuweb#2632.
-
Clarify GPUBufferUsage validation rules gpuweb/gpuweb#426.
-
GPURenderBundleEncoder shouldn't have a [[command_encoder]] gpuweb/gpuweb#1270.
-
scientific notation is not permitted in float literals (and A-5 problem) gpuweb/gpuweb#775.
-
Requiring integral-valued float literals to have .0 seems gratuitous gpuweb/gpuweb#739.
-
[wgsl] Support compile-time-constants (constexpr) gpuweb/gpuweb#1272.
-
Do we have a plan for non-32bit literals? gpuweb/gpuweb#1986.
-
wgsl: -0 is not a valid token gpuweb/gpuweb#2071.
-
Add scalar and vector constructors from abstract numeric types gpuweb/gpuweb#2719.
-
Consider using a more intuitive matrix notation gpuweb/gpuweb#2744.
-
dawn_wire crashes on normal ARM devices with SIGBUS (memory alignment issues) gpuweb/gpuweb#2739.
-
Literal unification gpuweb/gpuweb#2200.
-
Add identity constructors for matrices gpuweb/gpuweb#2704.
-
Disallow creating a single-layer 2d-array view of a 2d multisampled texture gpuweb/gpuweb#2715.
-
Clarify meaning of `dispatch()` args. gpuweb/gpuweb#2669.
-
Need to specify the order of dynamicOffsets in setBindGroup gpuweb/gpuweb#2392.
-
-
85 issues received 235 new comments.
-
[wgsl] Encoding for shader text needs to be specified gpuweb/gpuweb#565 (21 comments by @annevk, @ben-clayton, @kdashg, @litherum, @magcius, @mehmetoguzderin)
-
Combined limit for max vertex stage bind groups + vertex buffer bindings gpuweb/gpuweb#2749 (9 comments by @kainino0x, @Kangz, @litherum)
-
Concerns around error label and message gpuweb/gpuweb#2633 (9 comments by @Kangz, @litherum, @mehmetoguzderin, @toji)
-
Clarify undefined behavior of built in functions gpuweb/gpuweb#2765 (8 comments by @dneto0, @kdashg, @litherum, @teoxoy)
-
Support for arrays of textures gpuweb/gpuweb#822 (7 comments by @FishHawk, @gents83, @kainino0x, @Kangz, @kdashg, @litherum)
-
adapt line-ending comment to Unicode conventions gpuweb/gpuweb#2758 (6 comments by @dneto0, @mehmetoguzderin)
-
context-aware tokenization: Sometimes >> is better parsed as two copies of > (similar for ]] ) gpuweb/gpuweb#2092 (6 comments by @ben-clayton, @dj2, @dneto0, @kdashg)
-
my2DArrayTexture.createView() doesn't create a full view gpuweb/gpuweb#2684 (6 comments by @kainino0x, @Kangz, @toji)
-
The preferred format is not always storage-capable gpuweb/gpuweb#2748 (5 comments by @kainino0x, @Kangz)
-
GPUExternalTexture [[active_imports]] only works if there are RAFs gpuweb/gpuweb#2785 (5 comments by @kainino0x, @Kangz, @shaoboyan)
-
Potential parsing conflict with const literals and postfix expressions. gpuweb/gpuweb#2772 (5 comments by @dj2, @dneto0, @kdashg, @mehmetoguzderin)
-
Investigation : Push Constants gpuweb/gpuweb#75 (5 comments by @devshgraphicsprogramming, @Kangz, @sseanwang)
-
Allow copying from any canvas that toDataURL can be called on gpuweb/gpuweb#2373 (4 comments by @kainino0x, @shaoboyan, @toji)
-
Specify that texture_external sampling functions clamp coords to [0, 1] gpuweb/gpuweb#2766 (4 comments by @kainino0x, @Kangz, @kdashg)
-
Multisample resolve when base texture formats do not match gpuweb/gpuweb#2729 (4 comments by @austinEng, @kainino0x)
-
Consider allowing creation-time expressions or constants in switch case selectors gpuweb/gpuweb#2764 (4 comments by @alan-baker, @kdashg, @litherum, @teoxoy)
-
Are duplicate timestamp locations allowed on the same pass? gpuweb/gpuweb#2662 (4 comments by @haoxli, @kainino0x, @litherum)
-
Consider copyToTexture interactions with high-bit-depth source images gpuweb/gpuweb#1482 (3 comments by @kainino0x, @Kangz)
-
Change ArrayBuffer arguments to ArrayBuffer + srcOffset gpuweb/gpuweb#244 (3 comments by @domenic, @Kangz)
-
Internationalization Metadata for Human-readable Messages gpuweb/gpuweb#2780 (3 comments by @Kangz, @kdashg, @mehmetoguzderin)
-
Proposal: Support DP4a as WGSL built-in functions gpuweb/gpuweb#2677 (3 comments by @kdashg, @litherum, @mehmetoguzderin)
-
Missing Abstract Int/Float preconditions for Conversion Expressions gpuweb/gpuweb#2788 (2 comments by @alan-baker, @kdashg)
-
Consider allowing creation-time expressions for attributes gpuweb/gpuweb#2790 (2 comments by @kdashg, @litherum)
-
Consider making `binding_array` a reserved word gpuweb/gpuweb#2794 (2 comments by @jimblandy, @kdashg)
-
Metal primitive topology required for the shader compilation hint gpuweb/gpuweb#2574 (2 comments by @Kangz)
-
Dealing with CRLF and error line numbers gpuweb/gpuweb#2435 (2 comments by @kdashg, @mehmetoguzderin)
-
wgsl: lexical rule for blankspace should spell out the characters, not delegate to treesitter's \s gpuweb/gpuweb#2283 (2 comments by @mehmetoguzderin)
-
createRenderPipeline with preferred format is dependent on canvas gpuweb/gpuweb#2747 (2 comments by @kainino0x, @toji)
-
WebGPURenderingContext should support preserveDrawingBuffer: true gpuweb/gpuweb#2743 (2 comments by @kainino0x)
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520 (2 comments by @kainino0x, @Kangz)
-
Binary WGSL format in createShaderModule() WebAPI gpuweb/gpuweb#2571 (2 comments by @kvark, @magcius)
-
Add back constraint that DS views must have aspect "all" for rendering. gpuweb/gpuweb#2789 (1 comment by @Kangz)
-
Should objects expose the descriptor they were created with? gpuweb/gpuweb#1498 (1 comment by @Kangz)
-
Exposing GPU hardware information to web developers in WebGPU gpuweb/gpuweb#2195 (1 comment by @Kangz)
-
What should occur for inputs to `ldexp` that would produce a result that isn't representable as a `f32`? gpuweb/gpuweb#2631 (1 comment by @kdashg)
-
Dynamic texture selection is impossible for sample() calls gpuweb/gpuweb#2482 (1 comment by @kdashg)
-
call_statement: should only be able to call a function that yields void? gpuweb/gpuweb#1413 (1 comment by @kdashg)
-
Cannot upload/download to UMA storage buffer without an unnecessary copy and unnecessary memory use gpuweb/gpuweb#2388 (1 comment by @Kangz)
-
Clarify the vertex stage output / fragment stage input matching rules gpuweb/gpuweb#1953 (1 comment by @Jiawei-Shao)
-
Consider putting most of the writeTexture validation on the device timeline. gpuweb/gpuweb#2680 (1 comment by @Kangz)
-
scientific notation is not permitted in float literals (and A-5 problem) gpuweb/gpuweb#775 (1 comment by @dneto0)
-
Requiring integral-valued float literals to have .0 seems gratuitous gpuweb/gpuweb#739 (1 comment by @dneto0)
-
Does compositingAlphaMode affect using the canvas as an image source? gpuweb/gpuweb#1847 (1 comment by @kainino0x)
-
dawn_wire crashes on normal ARM devices with SIGBUS (memory alignment issues) gpuweb/gpuweb#2739 (1 comment by @hefeng-oppo)
-
Investigate possible solution for cheaper opaque canvases gpuweb/gpuweb#1988 (1 comment by @kainino0x)
-
Add scalar and vector constructors from abstract numeric types gpuweb/gpuweb#2719 (1 comment by @alan-baker)
-
-
51 pull requests opened.
-
Make IO-shareable types apply to user-defined IO only. gpuweb/gpuweb#2813, by @jimblandy.
-
Add validation for GPURender/ComputePassTimestampLocation gpuweb/gpuweb#2812, by @haoxli.
-
Remove restrictions on the canvas type for CopyExternalImageToTexture() gpuweb/gpuweb#2804, by @shaoboyan.
-
Specify how `getBindGroupLayout` handles out-of-bounds indices. gpuweb/gpuweb#2803, by @jimblandy.
-
Encoder mixin followups gpuweb/gpuweb#2799, by @kainino0x.
-
In type rule for boolean negation (`!`), use `T` in precondition. gpuweb/gpuweb#2797, by @jimblandy.
-
Make GPUExternalTexture sampling always use clamping address mode. gpuweb/gpuweb#2796, by @Kangz.
-
Clarity how to interpret GPURenderPassColorAttachment/clearValue gpuweb/gpuweb#2795, by @Jiawei-Shao.
-
Allow creation-time expression in case seletors; Add missings `@const`s gpuweb/gpuweb#2792, by @alan-baker.
-
Nits from #2751 gpuweb/gpuweb#2791, by @kainino0x.
-
Require RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/gpuweb#2786, by @Jiawei-Shao.
-
Clarify type and use case for object labels gpuweb/gpuweb#2784, by @toji.
-
[wgsl] Add some assert tags to the spec gpuweb/gpuweb#2783, by @dj2.
-
Adjust line-ending comment for unicode line break recommendations gpuweb/gpuweb#2782, by @dneto0.
-
Make blankspaces a section gpuweb/gpuweb#2781, by @mehmetoguzderin.
-
wgsl: Define "code point" by normative reference to Unicode gpuweb/gpuweb#2779, by @dneto0.
-
Clarify that evaluations to inf/nan may yield undef value gpuweb/gpuweb#2777, by @dneto0.
-
Adjust pipeline fragment output validation rules gpuweb/gpuweb#2775, by @shrekshao.
-
Don't reserve buffer, texture, in, out, input, output gpuweb/gpuweb#2774, by @dneto0.
-
Clarify conversions: floating, and out-of-range AbstractInt gpuweb/gpuweb#2773, by @dneto0.
-
Move preferred format queries to navigator.gpu gpuweb/gpuweb#2771, by @toji.
-
Adopt identical-to and further clarifications for both WGSL and WebGPU gpuweb/gpuweb#2767, by @mehmetoguzderin.
-
Set the default value of maxInterStageShaderComponents to 32 gpuweb/gpuweb#2763, by @litherum.
-
[wgsl] Split regex apart by segment. gpuweb/gpuweb#2762, by @dj2.
-
[wgsl] Add literal examples; fixup f16 hex floats regex. gpuweb/gpuweb#2761, by @dj2.
-
Add staticAssert gpuweb/gpuweb#2759, by @alan-baker.
-
[wgsl] Convert literal regex to grammar. gpuweb/gpuweb#2757, by @dj2.
-
Update logic for default TextureView dimension gpuweb/gpuweb#2755, by @toji.
-
Add WGSL grammar analyzer gpuweb/gpuweb#2754, by @dneto0.
-
docs: add the WebGPU logo to README.md gpuweb/gpuweb#2753, by @shuuji3.
-
Change hypothetical extensions example gpuweb/gpuweb#2752, by @alan-baker.
-
Finish reworking encoder mixins gpuweb/gpuweb#2751, by @kainino0x.
-
Expose different error types directly instead of via polymorphism gpuweb/gpuweb#2750, by @kainino0x.
-
[editorial] Change vast majority of must language to link to shader-creation error gpuweb/gpuweb#2746, by @alan-baker.
-
Make matrix notation more intuitive gpuweb/gpuweb#2745, by @teoxoy.
-
Rewrite the function aliasing restrictions gpuweb/gpuweb#2742, by @alan-baker.
-
WGSL: Add dot4U8Packed and dot4I8Packed as new built-in functions gpuweb/gpuweb#2738, by @Jiawei-Shao.
-
Fix parse conflict between function call and type constructor gpuweb/gpuweb#2737, by @dneto0.
-
WGSL: Don't reserve words that are types in other languages gpuweb/gpuweb#2736, by @dneto0.
-
Describe WGSL's context-aware tokenization gpuweb/gpuweb#2735, by @dneto0.
-
WGSL Internationalization Adjustment for Code Points gpuweb/gpuweb#2734, by @mehmetoguzderin.
-
[wgsl] Link CTS from spec. gpuweb/gpuweb#2733, by @dj2.
-
[wgsl] Add some railroad diagrams gpuweb/gpuweb#2731, by @dj2.
-
[wgsl] Clarify the continue text. gpuweb/gpuweb#2728, by @dj2.
-
[wgsl] Match any and all formatting in logical built-in table. gpuweb/gpuweb#2727, by @dj2.
-
[wgsl] Minor formatting and language changes. gpuweb/gpuweb#2726, by @dj2.
-
Fix typo gpuweb/gpuweb#2725, by @OmarShehata.
-
[wgsl] Allow continue to be called within a while loop. gpuweb/gpuweb#2724, by @dj2.
-
Add comments to the WGSL grammar extractor gpuweb/gpuweb#2723, by @dneto0.
-
Change integer texture parameters to unsigned gpuweb/gpuweb#2722, by @alan-baker.
-
wgsl: Allow increment/decrement statements in for_init gpuweb/gpuweb#2721, by @jrprice.
-
-
37 pull requests merged.
-
Remove restrictions on the canvas type for CopyExternalImageToTexture() gpuweb/gpuweb#2804.
-
Clarify type and use case for object labels gpuweb/gpuweb#2784.
-
Specify how `getBindGroupLayout` handles out-of-bounds indices. gpuweb/gpuweb#2803.
-
Encoder mixin followups gpuweb/gpuweb#2799.
-
Don't reserve buffer, texture, in, out, input, output gpuweb/gpuweb#2774.
-
Require RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/gpuweb#2786.
-
Nits from #2751 gpuweb/gpuweb#2791.
-
Adjust line-ending comment for unicode line break recommendations gpuweb/gpuweb#2782.
-
Make blankspaces a section gpuweb/gpuweb#2781.
-
wgsl: Define "code point" by normative reference to Unicode gpuweb/gpuweb#2779.
-
Adopt identical-to and further clarifications for both WGSL and WebGPU gpuweb/gpuweb#2767.
-
Update logic for default TextureView dimension gpuweb/gpuweb#2755.
-
[wgsl] Add literal examples; fixup f16 hex floats regex. gpuweb/gpuweb#2761.
-
docs: add the WebGPU logo to README.md gpuweb/gpuweb#2753.
-
Finish reworking encoder mixins gpuweb/gpuweb#2751.
-
[wgsl] Minor formatting and language changes. gpuweb/gpuweb#2726.
-
WGSL Internationalization Adjustment for Code Points gpuweb/gpuweb#2734.
-
Change hypothetical extensions example gpuweb/gpuweb#2752.
-
Add creation-time expressions gpuweb/gpuweb#2668.
-
Make matrix notation more intuitive gpuweb/gpuweb#2745.
-
Use concise forms for entry point stage attributes. gpuweb/gpuweb#2652.
-
FP16 Extension against spec gpuweb/gpuweb#2696.
-
Rewrite the function aliasing restrictions gpuweb/gpuweb#2742.
-
WGSL: Don't reserve words that are types in other languages gpuweb/gpuweb#2736.
-
Fix parse conflict between function call and type constructor gpuweb/gpuweb#2737.
-
Correct the definition of fragment-stage position builtin gpuweb/gpuweb#2625.
-
Add identity constructors for matrix types gpuweb/gpuweb#2714.
-
wgsl: Allow increment/decrement statements in for_init gpuweb/gpuweb#2721.
-
[wgsl] Clarify the continue text. gpuweb/gpuweb#2728.
-
[wgsl] Link CTS from spec. gpuweb/gpuweb#2733.
-
Require views on 2d multisampled textures to be 2d gpuweb/gpuweb#2716.
-
[wgsl] Match any and all formatting in logical built-in table. gpuweb/gpuweb#2727.
-
Add comments to the WGSL grammar extractor gpuweb/gpuweb#2723.
-
Rename dispatch -> dispatchWorkgroups gpuweb/gpuweb#2689.
-
Better specify the order of dynamicOffsets in setBindGroup gpuweb/gpuweb#2710.
-
Fix typo gpuweb/gpuweb#2725.
-
[wgsl] Allow continue to be called within a while loop. gpuweb/gpuweb#2724.
-
-
8 pull requests closed.
-
Adjust pipeline fragment output validation rules gpuweb/gpuweb#2775.
-
Expose fast-math gpuweb/gpuweb#2077.
-
reworking uniformity analysis PR gpuweb/gpuweb#2487.
-
[wgsl] Convert literal regex to grammar. gpuweb/gpuweb#2757.
-
FP16 Extension Document gpuweb/gpuweb#2647.
-
[wgsl] Add some railroad diagrams gpuweb/gpuweb#2731.
-
Int and float literals don't include leading minus gpuweb/gpuweb#1092.
-
Literal unification gpuweb/gpuweb#2221.
-
-
64 pull requests received 323 new comments.
-
FP16 Extension against spec gpuweb/gpuweb#2696 (30 comments by @alan-baker, @dneto0, @jiangzhaoming, @kdashg, @tidoust)
-
[wgsl] Minor formatting and language changes. gpuweb/gpuweb#2726 (29 comments by @alan-baker, @dj2, @dneto0, @kdashg)
-
[wgsl] Split regex apart by segment. gpuweb/gpuweb#2762 (26 comments by @dj2, @dneto0, @mehmetoguzderin, @teoxoy)
-
[editorial] Change vast majority of must language to link to shader-creation error gpuweb/gpuweb#2746 (19 comments by @alan-baker, @dj2, @dneto0, @kdashg)
-
WGSL Internationalization Adjustment for Code Points gpuweb/gpuweb#2734 (19 comments by @dj2, @dneto0, @kdashg, @litherum, @mehmetoguzderin)
-
Clarity how to interpret GPURenderPassColorAttachment/clearValue gpuweb/gpuweb#2795 (18 comments by @Jiawei-Shao, @kainino0x, @Kangz)
-
Add GPUAdaperInfo to the spec gpuweb/gpuweb#2660 (18 comments by @kainino0x, @Kangz, @rconde01, @toji)
-
Adopt identical-to and further clarifications for both WGSL and WebGPU gpuweb/gpuweb#2767 (8 comments by @dneto0, @mehmetoguzderin, @toji)
-
Add staticAssert gpuweb/gpuweb#2759 (7 comments by @alan-baker, @ben-clayton, @dj2, @dneto0, @kdashg)
-
Loosen pipeline fragment output validation gpuweb/gpuweb#1918 (6 comments by @austinEng, @Kangz, @shrekshao)
-
WGSL: Explain parameterized type rules more explicitly. gpuweb/gpuweb#2682 (6 comments by @alan-baker, @jimblandy, @kdashg)
-
[wgsl] Add literal examples; fixup f16 hex floats regex. gpuweb/gpuweb#2761 (6 comments by @dj2, @dneto0, @kdashg, @mehmetoguzderin)
-
Rely on canvas size instead of allowing size to be configured gpuweb/gpuweb#2416 (6 comments by @annevk, @kainino0x)
-
[wgsl] Convert literal regex to grammar. gpuweb/gpuweb#2757 (5 comments by @dj2, @dneto0, @mehmetoguzderin)
-
Make matrix notation more intuitive gpuweb/gpuweb#2745 (5 comments by @dneto0, @kdashg, @teoxoy, @w3cbot)
-
Describe WGSL's context-aware tokenization gpuweb/gpuweb#2735 (5 comments by @alan-baker, @dneto0, @mehmetoguzderin)
-
Expose different error types directly instead of via polymorphism gpuweb/gpuweb#2750 (4 comments by @kainino0x, @toji)
-
Allow creation-time expression in case seletors; Add missings `@const`s gpuweb/gpuweb#2792 (4 comments by @alan-baker, @kdashg)
-
Add comments to the WGSL grammar extractor gpuweb/gpuweb#2723 (4 comments by @dneto0, @mehmetoguzderin)
-
Clarify that evaluations to inf/nan may yield undef value gpuweb/gpuweb#2777 (3 comments by @dneto0, @kdashg)
-
Adjust line-ending comment for unicode line break recommendations gpuweb/gpuweb#2782 (3 comments by @dneto0)
-
Change integer texture parameters to unsigned gpuweb/gpuweb#2722 (3 comments by @alan-baker, @kdashg, @litherum)
-
Adjust pipeline fragment output validation rules gpuweb/gpuweb#2775 (2 comments by @Kangz, @shrekshao)
-
Add validation for GPURender/ComputePassTimestampLocation gpuweb/gpuweb#2812 (2 comments by @kainino0x)
-
Set the default value of maxInterStageShaderComponents to 32 gpuweb/gpuweb#2763 (2 comments by @Kangz, @litherum)
-
WGSL: Don't reserve words that are types in other languages gpuweb/gpuweb#2736 (2 comments by @dneto0, @litherum)
-
Better specify the order of dynamicOffsets in setBindGroup gpuweb/gpuweb#2710 (2 comments by @kainino0x, @toji)
-
Add proposals/ directory and pipeline statistics extension doc gpuweb/gpuweb#2301 (2 comments by @Kangz)
-
Specify how `getBindGroupLayout` handles out-of-bounds indices. gpuweb/gpuweb#2803 (1 comment by @Kangz)
-
Make GPUExternalTexture sampling always use clamping address mode. gpuweb/gpuweb#2796 (1 comment by @kainino0x)
-
Clarify conversions: floating, and out-of-range AbstractInt gpuweb/gpuweb#2773 (1 comment by @kdashg)
-
Move all canvas configure validation to getCurrentTexture gpuweb/gpuweb#2536 (1 comment by @kainino0x)
-
Require RENDER_ATTACHMENT usage when creating multisampled textures gpuweb/gpuweb#2786 (1 comment by @kainino0x)
-
wgsl: Define "code point" by normative reference to Unicode gpuweb/gpuweb#2779 (1 comment by @dneto0)
-
WGSL: Add dot4U8Packed and dot4I8Packed as new built-in functions gpuweb/gpuweb#2738 (1 comment by @kdashg)
-
gpuweb/cts
-
104 issues opened.
-
correctlyRoundedThreshold does not correctly handling some cases gpuweb/cts#1332, by @zoddicus.
-
wgsl: More Abstract number tests gpuweb/cts#1323, by @dj2.
-
Shorten maths test names gpuweb/cts#1316, by @zoddicus.
-
wgsl: Add abstract float/int tests to existing specs gpuweb/cts#1297, by @dj2.
-
wgsl: execution test for workgroupBarrier builtin gpuweb/cts#1295, by @dj2.
-
wgsl: execution test for storageBarrier builtin gpuweb/cts#1294, by @dj2.
-
wgsl: execution test for unpack2x16float builtin gpuweb/cts#1293, by @dj2.
-
wgsl: execution test for unpack2x16unorm builtin gpuweb/cts#1292, by @dj2.
-
wgsl: execution test for unpack2x16snorm gpuweb/cts#1291, by @dj2.
-
wgsl: execution test for unpack4x8unorm builtin gpuweb/cts#1290, by @dj2.
-
wgsl: execution test for unpack4x8snorm builtin gpuweb/cts#1289, by @dj2.
-
wgsl: execution test for pack2x16float builtin gpuweb/cts#1288, by @dj2.
-
wgsl: execution test for pack2x16unorm builtin gpuweb/cts#1287, by @dj2.
-
wgsl: execution test for pack2x16snorm gpuweb/cts#1286, by @dj2.
-
wgsl: execution test for pack4x8unorm builtin gpuweb/cts#1285, by @dj2.
-
wgsl: execution test for pack4x8snorm builtin gpuweb/cts#1284, by @dj2.
-
wgsl: execution test for atomicCompareExchangeWeak builtin gpuweb/cts#1283, by @dj2.
-
wgsl: execution test for atomicExchange builtin gpuweb/cts#1282, by @dj2.
-
wgsl: execution test for atomicXor builtin gpuweb/cts#1281, by @dj2.
-
wgsl: execution test for atomicOr builtin gpuweb/cts#1280, by @dj2.
-
wgsl: execution test for atomicAnd builtin gpuweb/cts#1279, by @dj2.
-
wgsl: execution test for atomicMin builtin gpuweb/cts#1278, by @dj2.
-
wgsl: execution test for atomicMax builtin gpuweb/cts#1277, by @dj2.
-
wgsl: execution test for atomicSub builtin gpuweb/cts#1276, by @dj2.
-
wgsl: execution test for atomicAdd builtin gpuweb/cts#1275, by @dj2.
-
wgsl: execution test for atomicStore builtin gpuweb/cts#1274, by @dj2.
-
wgsl: execution test for atomicLoad builtin gpuweb/cts#1273, by @dj2.
-
wgsl: execution test for textureStore builtin gpuweb/cts#1272, by @dj2.
-
wgsl: execution test for textureSampleLevel builtin gpuweb/cts#1271, by @dj2.
-
wgsl: execution test for textureSampleGrad builtin gpuweb/cts#1270, by @dj2.
-
wgsl: execution test for textureSampleCompareLevel builtin gpuweb/cts#1269, by @dj2.
-
wgsl: execution test for textureSampleCompare builtin gpuweb/cts#1268, by @dj2.
-
wgsl: execution test for textureSampleBias builtin gpuweb/cts#1267, by @dj2.
-
wgsl: execution test for textureSample builtin gpuweb/cts#1266, by @dj2.
-
wgsl: execution test for textureNumSamples builtin gpuweb/cts#1265, by @dj2.
-
wgsl: execution test for textureNumLevels gpuweb/cts#1264, by @dj2.
-
wgsl: execution test for textureNumLayers gpuweb/cts#1263, by @dj2.
-
wgsl: execution test for textureLoad gpuweb/cts#1262, by @dj2.
-
wgsl: execution test for textureGatherCompare builtin gpuweb/cts#1260, by @dj2.
-
wgsl: execution tests for textureGather builtin gpuweb/cts#1259, by @dj2.
-
wgsl: execution tests for textureDimensions builtin gpuweb/cts#1258, by @dj2.
-
wgsl: execution test for fwidthFine builtin gpuweb/cts#1257, by @dj2.
-
wgsl: execution test for fwidthCoarse builtin gpuweb/cts#1256, by @dj2.
-
wgsl: execution test for fwidth builtin gpuweb/cts#1255, by @dj2.
-
wgsl: execution test for dpdyFine builtin gpuweb/cts#1254, by @dj2.
-
wgsl: execution test for dpdyCoarse builtin gpuweb/cts#1253, by @dj2.
-
wgsl: execution test for dpdy builtin gpuweb/cts#1252, by @dj2.
-
wgsl: execution test for dpdxFine builtin gpuweb/cts#1251, by @dj2.
-
wgsl: execution test for dpdxCoarse builtin gpuweb/cts#1250, by @dj2.
-
wgsl: Add execution test for dpdx builtin gpuweb/cts#1249, by @dj2.
-
wgsl: f16 built-in execution tests gpuweb/cts#1248, by @dj2.
-
wgsl: f32 execution test for `dot` builtin gpuweb/cts#1247, by @dj2.
-
wgsl: f32 execution test for `transpose` builtin gpuweb/cts#1246, by @dj2.
-
wgsl: f32 execution test for `determinant` builtin gpuweb/cts#1245, by @dj2.
-
wgsl: execution test for shift* methods gpuweb/cts#1244, by @dj2.
-
wgsl: f32 execution test for `trunc` builtin gpuweb/cts#1243, by @dj2.
-
wgsl : f32 execution test for `tanh` builtin. gpuweb/cts#1242, by @dj2.
-
wgsl: f32 execution test for `tan` builtin gpuweb/cts#1241, by @dj2.
-
wgsl: f32 execution test for `step` builtin gpuweb/cts#1240, by @dj2.
-
wgsl: f32 execution test for `sqrt` builtin gpuweb/cts#1239, by @dj2.
-
wgsl: f32 execution test for `smoothstep` builtin gpuweb/cts#1238, by @dj2.
-
wgsl: f32 execution test for `sinh` builtin gpuweb/cts#1237, by @dj2.
-
wgsl: f32 execution tests for `sign` builtin gpuweb/cts#1236, by @dj2.
-
wgsl: f32 execution test for round builtin gpuweb/cts#1235, by @dj2.
-
wgsl: f32 execution tests for `refract` builtin gpuweb/cts#1234, by @dj2.
-
wgsl: f32 execution tests for `reflect` builtin gpuweb/cts#1233, by @dj2.
-
wgsl: f32 execution tests for `radians` builtin gpuweb/cts#1232, by @dj2.
-
wgsl: f32 execution tests for `pow` builtin gpuweb/cts#1231, by @dj2.
-
wgsl: f32 execution tests for `normalize` builtin gpuweb/cts#1230, by @dj2.
-
wgsl: execution test for modf for vector f16 gpuweb/cts#1229, by @dj2.
-
wgsl: execution test for modf for vector f32 gpuweb/cts#1228, by @dj2.
-
wgsl: execution test for modf for scalar f16 gpuweb/cts#1227, by @dj2.
-
wgsl: f32 execution tests for modf gpuweb/cts#1226, by @dj2.
-
wgsl: f32 execution tests for `mix` builtin gpuweb/cts#1225, by @dj2.
-
wgsl: f32 execution tests for `degrees` builtin gpuweb/cts#1224, by @dj2.
-
wgsl: f32 execution tests for `atanh` builtin gpuweb/cts#1223, by @dj2.
-
wgsl: f32 execution tests for `asinh` builtin gpuweb/cts#1222, by @dj2.
-
wgsl: f32 execution tests for `acosh` builtin gpuweb/cts#1221, by @dj2.
-
wgsl: f32 execution tests for `distance` builtin gpuweb/cts#1217, by @dj2.
-
wgsl: f32 execution tests for `length` builtin gpuweb/cts#1216, by @dj2.
-
wgsl: f32 execution tests for `asin` builtin gpuweb/cts#1213, by @dj2.
-
wgsl: f32 execution tests for `acos` builtin gpuweb/cts#1212, by @dj2.
-
wgsl: execution test for frexp for vector f16 gpuweb/cts#1211, by @dj2.
-
wgsl: execution test for frexp for vector f32 gpuweb/cts#1210, by @dj2.
-
wgsl: execution test for frexp for scalar f16 gpuweb/cts#1209, by @dj2.
-
wgsl: execution test for frexp for scalar f32 gpuweb/cts#1208, by @dj2.
-
wgsl: f32 execution tests for `fma` builtin gpuweb/cts#1207, by @dj2.
-
wgsl: f32 execution tests for `faceForward` builtin gpuweb/cts#1206, by @dj2.
-
wgsl: f32 execution tests for `exp2` builtin gpuweb/cts#1205, by @dj2.
-
wgsl: f32 execution tests for `exp` builtin gpuweb/cts#1204, by @dj2.
-
wgsl: f32 execution tests for `cross` gpuweb/cts#1203, by @dj2.
-
wgsl: f32 execution tests for `cosh` gpuweb/cts#1202, by @dj2.
-
wgsl: execution test for `arrayLength` builtin gpuweb/cts#1201, by @dj2.
-
wgsl: execution test quantizeToF16 gpuweb/cts#1200, by @dj2.
-
wgsl: literal parsing execution tests gpuweb/cts#1192, by @dj2.
-
Identifier tests with unicode names failing in standalone viewer gpuweb/cts#1191, by @dj2.
-
Support f16 as a type gpuweb/cts#1187, by @dj2.
-
should test use of implicit derivatives in point rasterization mode gpuweb/cts#1182, by @dneto0.
-
textureContentIsOKByT2B is slow gpuweb/cts#1162, by @Kangz.
-
wgsl: blankspace definition has changed gpuweb/cts#1159, by @dneto0.
-
Change how ULP threshholds are calculated gpuweb/cts#1148, by @zoddicus.
-
Issue with interpolate wgsl test gpuweb/cts#1135, by @dj2.
-
Extend the ShaderModuleEntryPoint tests with Unicode gpuweb/cts#1130, by @Kangz.
-
-
10 issues closed.
-
Shorten maths test names gpuweb/cts#1316.
-
wgsl: Add tests for builtin `exp` gpuweb/cts#1125.
-
Change how ULP threshholds are calculated gpuweb/cts#1148.
-
Identifier tests with unicode names failing in standalone viewer gpuweb/cts#1191.
-
Issue with interpolate wgsl test gpuweb/cts#1135.
-
wgsl: validate identifier token pattern gpuweb/cts#748.
-
wgsl: blankspace definition has changed gpuweb/cts#1159.
-
How to make sure a WGSL test is failing for the right reasons? gpuweb/cts#125.
-
val: queue,destroyed,buffer gpuweb/cts#879.
-
val: attachment_compatibility for read-only depth/stencil gpuweb/cts#927.
-
-
14 issues received 19 new comments.
-
Identifier tests with unicode names failing in standalone viewer gpuweb/cts#1191 (2 comments by @dj2)
-
should test use of implicit derivatives in point rasterization mode gpuweb/cts#1182 (1 comment by @dneto0)
-
-
105 pull requests opened.
-
Don't block on subcases until the end of a testcase gpuweb/cts#1335, by @austinEng.
-
Roll @webgpu/types to 0.1.15 gpuweb/cts#1334, by @toji.
-
Move test device selection to an optional before clause gpuweb/cts#1333, by @austinEng.
-
wgsl: Implement `exp2` f32 tests gpuweb/cts#1331, by @zoddicus.
-
Add comprehensive memory model tests gpuweb/cts#1330, by @reeselevine.
-
wgsl: Add stub tests for round builtin gpuweb/cts#1329, by @dj2.
-
wgsl: Stub tests for refract builtin. gpuweb/cts#1328, by @dj2.
-
wgsl: Stub tests for reflect builtin gpuweb/cts#1327, by @dj2.
-
wgsl: Stub test for radians builtin. gpuweb/cts#1326, by @dj2.
-
wgsl: Stub test for `pow` builtin. gpuweb/cts#1325, by @dj2.
-
wgsl: Stub normalize builtin test gpuweb/cts#1324, by @dj2.
-
wgsl: Stub tests for modf builtin. gpuweb/cts#1322, by @dj2.
-
wgsl: Stub tests for mix builtin. gpuweb/cts#1321, by @dj2.
-
wgsl: Stub tests for degrees builtin. gpuweb/cts#1320, by @dj2.
-
depth_clip_clamp: Fix validation error for stencil-less formats gpuweb/cts#1319, by @Kangz.
-
Allow 1 ULP difference for copyToTexture ImageBitmap tests gpuweb/cts#1318, by @Kangz.
-
Remove redundant information from unittest names gpuweb/cts#1317, by @zoddicus.
-
wgsl: Implement f32 `exp` test gpuweb/cts#1315, by @zoddicus.
-
wgsl: stub tests for atanh builtin. gpuweb/cts#1314, by @dj2.
-
wgsl: Stub asinh builtin tests gpuweb/cts#1313, by @dj2.
-
wgsl: Stub acosh builtin tests. gpuweb/cts#1312, by @dj2.
-
wgsl: stub distance builtin tests gpuweb/cts#1311, by @dj2.
-
wgsl: stub length builtin tests gpuweb/cts#1310, by @dj2.
-
wgsl: Stub asin builtin tests gpuweb/cts#1309, by @dj2.
-
Sync reserved words to spec. gpuweb/cts#1308, by @dj2.
-
Add positive case into invariant duplicate test. gpuweb/cts#1307, by @dj2.
-
wgsl: Stub acos builtin tests gpuweb/cts#1306, by @dj2.
-
wgsl: Stub frexp test gpuweb/cts#1305, by @dj2.
-
wgsl: Stub fma builtin tests. gpuweb/cts#1304, by @dj2.
-
wgsl: Stub faceFoward tests. gpuweb/cts#1303, by @dj2.
-
wgsl: Stub exp2 builtin tests. gpuweb/cts#1302, by @dj2.
-
wgsl: Add stub for exp builtin gpuweb/cts#1301, by @dj2.
-
wgsl: Stub out cross builtin tests gpuweb/cts#1300, by @dj2.
-
wgsl: Add unimplemented stub for cosh builtin. gpuweb/cts#1299, by @dj2.
-
wgsl: Add unimplemented quantizeToF16 spec gpuweb/cts#1298, by @dj2.
-
Sync WGSL spec descriptions to spec text. gpuweb/cts#1296, by @dj2.
-
Fix TypeScript errors with build-out-node. gpuweb/cts#1261, by @Kangz.
-
Remove old builtin combined function file. gpuweb/cts#1220, by @dj2.
-
Remove uniqueId and plan_autogen from WGSL tests. gpuweb/cts#1219, by @dj2.
-
Implement OneULP and related ULP functions gpuweb/cts#1218, by @zoddicus.
-
src/webgpu/util/texture/texel_data: Use `,` for member delimiters gpuweb/cts#1215, by @ben-clayton.
-
cmdline: Add --list-unimplemented gpuweb/cts#1214, by @ben-clayton.
-
Rename generic WGSL shader_io test case. gpuweb/cts#1199, by @dj2.
-
shader/execution: Reduce `atan2` cases gpuweb/cts#1198, by @ben-clayton.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part II gpuweb/cts#1197, by @Jiawei-Shao.
-
Test invariant can only be applied once. gpuweb/cts#1196, by @dj2.
-
Add more interpolate tests. gpuweb/cts#1195, by @dj2.
-
Add a test with a Unicode character which would normalize gpuweb/cts#1194, by @dj2.
-
Add literal parsing tests gpuweb/cts#1193, by @dj2.
-
Add a quiet option to the test runner. gpuweb/cts#1190, by @dj2.
-
Extend WGSL interpolate tests. gpuweb/cts#1189, by @dj2.
-
Update WGSL identifier tests. gpuweb/cts#1188, by @dj2.
-
Update blankspace tests to match spec. gpuweb/cts#1186, by @dj2.
-
Update comment ending characters. gpuweb/cts#1185, by @dj2.
-
ShaderValidationTest: remove string overload gpuweb/cts#1184, by @dj2.
-
Split WGSL test files gpuweb/cts#1183, by @dj2.
-
Cleanup old WGSL test files. gpuweb/cts#1181, by @dj2.
-
Minor comment fixes gpuweb/cts#1180, by @dj2.
-
Take new createView() dimension default logic into account gpuweb/cts#1179, by @toji.
-
Explicitly destroy devices created by requestDevice_limits gpuweb/cts#1178, by @toji.
-
Use subcases in subresources,buffer_usage_in_render_pass gpuweb/cts#1177, by @Jiawei-Shao.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part I gpuweb/cts#1176, by @Jiawei-Shao.
-
Remove CopyExternalImageToTexture:destination_texture,dimension:* gpuweb/cts#1175, by @Jiawei-Shao.
-
Update pipeline output targets validation with different writeMask gpuweb/cts#1174, by @shrekshao.
-
Add validation,resource_usages,texture,in_render_misc:* - Part IV gpuweb/cts#1173, by @Jiawei-Shao.
-
Pipeline gathering subcase results gpuweb/cts#1172, by @austinEng.
-
Defer subcase finalization until the end of the testcase gpuweb/cts#1170, by @austinEng.
-
Add an option to run subcases in parallel gpuweb/cts#1169, by @austinEng.
-
Fix a bug in validation,createTexture,texture_usage:dimension=undef gpuweb/cts#1168, by @Jiawei-Shao.
-
Add validation,resource_usages,texture,in_render_misc:* - Part III gpuweb/cts#1167, by @Jiawei-Shao.
-
Validation test for GPURenderPassDepthStencilAttachment usage gpuweb/cts#1166, by @shrekshao.
-
Add presubmit check for long test names gpuweb/cts#1165, by @austinEng.
-
Reduce the length of subresource test names gpuweb/cts#1164, by @Kangz.
-
zero_init: add constant zero to prevent driver unrolling loops. gpuweb/cts#1163, by @Kangz.
-
Add validation,resource_usages,texture,in_render_misc:* - Part II gpuweb/cts#1161, by @Jiawei-Shao.
-
Add sparse color attachments operation tests gpuweb/cts#1160, by @shrekshao.
-
Make things work on Node 12.x, and test it on presubmit. gpuweb/cts#1158, by @kainino0x.
-
Add validation,resource_usages,texture,in_render_misc:* - Part I gpuweb/cts#1157, by @Jiawei-Shao.
-
Only 2D textures can be marked as render targets gpuweb/cts#1155, by @litherum.
-
operation,image_copy: Add support for depth16unorm gpuweb/cts#1154, by @Kangz.
-
depth_clip_clamp: only set stencilLoadOp and friends when needed. gpuweb/cts#1153, by @Kangz.
-
depth_compare_func: Don't use 0.5 as depth16unorm rounds either way gpuweb/cts#1152, by @Kangz.
-
image_copy,mip_levels: reduce the mip level count. gpuweb/cts#1151, by @Kangz.
-
Add validation,resource_usages,texture,in_render_common:* - Part V gpuweb/cts#1150, by @Jiawei-Shao.
-
Sparse color attachments validation tests gpuweb/cts#1149, by @shrekshao.
-
Factor float numeric range generation into common location gpuweb/cts#1147, by @zoddicus.
-
Fix tests trying to copy aspects of depth-stencil textures. gpuweb/cts#1146, by @Kangz.
-
Fix validation errors with depth/StencilLoad/StoreOp set when not all… gpuweb/cts#1145, by @Kangz.
-
Shorten test and param lengths in resource_usages,texture,in_render_common:* gpuweb/cts#1144, by @austinEng.
-
Fix infinite recursion in CopyExternalImageToTexture.spec.ts gpuweb/cts#1143, by @austinEng.
-
Update DevicePool to provide multiple devices gpuweb/cts#1142, by @haoxli.
-
Use globalThis instead of self. gpuweb/cts#1141, by @Kangz.
-
val,buffer,create: Buffer sizes don't have to be multiple of 4. gpuweb/cts#1140, by @Kangz.
-
image_copy: Ensure bufferOffset % 4 == 0 for depth-stencil. gpuweb/cts#1139, by @Kangz.
-
Don't require offset % 4 == 0 for depth-stencil <-> buffer copies gpuweb/cts#1138, by @Kangz.
-
Fix test of BGL binding limits to correctly check limits gpuweb/cts#1137, by @Kangz.
-
Fix DSAttachment load/store ops being specified when disallowed. gpuweb/cts#1136, by @Kangz.
-
val/copyT2T,format_compat: Don't assume compressed is always 4x4 gpuweb/cts#1134, by @Kangz.
-
Fix a validation error caused by WriteTexture bytesPerRow: t.width gpuweb/cts#1133, by @Kangz.
-
Add validation,resource_usages,texture,in_render_common:* - Part IV gpuweb/cts#1132, by @Jiawei-Shao.
-
Add unicode tests for entryPoint matching. gpuweb/cts#1131, by @Kangz.
-
writeBuffer: an unaligned bufferOffset is a validation error. gpuweb/cts#1129, by @Kangz.
-
validation,vertex_state: Only check valid attribute @locations gpuweb/cts#1128, by @Kangz.
-
Destroyed buffer submitting validation review and impl (#879) gpuweb/cts#1127, by @jiangzhaoming.
-
-
97 pull requests merged.
-
Add sparse color attachments operation tests gpuweb/cts#1160.
-
Roll @webgpu/types to 0.1.15 gpuweb/cts#1334.
-
wgsl: Implement `exp2` f32 tests gpuweb/cts#1331.
-
wgsl: Stub normalize builtin test gpuweb/cts#1324.
-
wgsl: Stub test for `pow` builtin. gpuweb/cts#1325.
-
wgsl: Stub test for radians builtin. gpuweb/cts#1326.
-
wgsl: Stub tests for reflect builtin gpuweb/cts#1327.
-
wgsl: Stub tests for refract builtin. gpuweb/cts#1328.
-
wgsl: Add stub tests for round builtin gpuweb/cts#1329.
-
wgsl: Stub tests for modf builtin. gpuweb/cts#1322.
-
wgsl: Stub tests for mix builtin. gpuweb/cts#1321.
-
depth_clip_clamp: Fix validation error for stencil-less formats gpuweb/cts#1319.
-
Allow 1 ULP difference for copyToTexture ImageBitmap tests gpuweb/cts#1318.
-
Add literal parsing tests gpuweb/cts#1193.
-
wgsl: Stub tests for degrees builtin. gpuweb/cts#1320.
-
wgsl: Implement f32 `exp` test gpuweb/cts#1315.
-
Remove redundant information from unittest names gpuweb/cts#1317.
-
wgsl: Stub acosh builtin tests. gpuweb/cts#1312.
-
wgsl: stub distance builtin tests gpuweb/cts#1311.
-
wgsl: stub length builtin tests gpuweb/cts#1310.
-
wgsl: stub tests for atanh builtin. gpuweb/cts#1314.
-
wgsl: Stub asinh builtin tests gpuweb/cts#1313.
-
Implement OneULP and related ULP functions gpuweb/cts#1218.
-
wgsl: Stub asin builtin tests gpuweb/cts#1309.
-
Sync reserved words to spec. gpuweb/cts#1308.
-
Add positive case into invariant duplicate test. gpuweb/cts#1307.
-
Add more interpolate tests. gpuweb/cts#1195.
-
wgsl: Stub frexp test gpuweb/cts#1305.
-
wgsl: Stub acos builtin tests gpuweb/cts#1306.
-
Test invariant can only be applied once. gpuweb/cts#1196.
-
wgsl: Stub faceFoward tests. gpuweb/cts#1303.
-
wgsl: Stub fma builtin tests. gpuweb/cts#1304.
-
wgsl: Stub exp2 builtin tests. gpuweb/cts#1302.
-
wgsl: Add stub for exp builtin gpuweb/cts#1301.
-
wgsl: Stub out cross builtin tests gpuweb/cts#1300.
-
wgsl: Add unimplemented stub for cosh builtin. gpuweb/cts#1299.
-
cmdline: Add --list-unimplemented gpuweb/cts#1214.
-
wgsl: Add unimplemented quantizeToF16 spec gpuweb/cts#1298.
-
Sync WGSL spec descriptions to spec text. gpuweb/cts#1296.
-
Fix TypeScript errors with build-out-node. gpuweb/cts#1261.
-
Remove old builtin combined function file. gpuweb/cts#1220.
-
src/webgpu/util/texture/texel_data: Use `,` for member delimiters gpuweb/cts#1215.
-
Add a test with a Unicode character which would normalize gpuweb/cts#1194.
-
Remove uniqueId and plan_autogen from WGSL tests. gpuweb/cts#1219.
-
Add a quiet option to the test runner. gpuweb/cts#1190.
-
Rename generic WGSL shader_io test case. gpuweb/cts#1199.
-
shader/execution: Reduce `atan2` cases gpuweb/cts#1198.
-
Extend WGSL interpolate tests. gpuweb/cts#1189.
-
Update WGSL identifier tests. gpuweb/cts#1188.
-
Update comment ending characters. gpuweb/cts#1185.
-
Update blankspace tests to match spec. gpuweb/cts#1186.
-
ShaderValidationTest: remove string overload gpuweb/cts#1184.
-
Split WGSL test files gpuweb/cts#1183.
-
Minor comment fixes gpuweb/cts#1180.
-
Cleanup old WGSL test files. gpuweb/cts#1181.
-
Take new createView() dimension default logic into account gpuweb/cts#1179.
-
Explicitly destroy devices created by requestDevice_limits gpuweb/cts#1178.
-
Use subcases in subresources,buffer_usage_in_render_pass gpuweb/cts#1177.
-
Add api,validation,resource_usages,buffer,in_pass_encoder:* - Part I gpuweb/cts#1176.
-
Remove CopyExternalImageToTexture:destination_texture,dimension:* gpuweb/cts#1175.
-
Add validation,resource_usages,texture,in_render_misc:* - Part IV gpuweb/cts#1173.
-
Factor float numeric range generation into common location gpuweb/cts#1147.
-
Sparse color attachments validation tests gpuweb/cts#1149.
-
Fix a bug in validation,createTexture,texture_usage:dimension=undef gpuweb/cts#1168.
-
Add validation,resource_usages,texture,in_render_misc:* - Part III gpuweb/cts#1167.
-
Validation test for GPURenderPassDepthStencilAttachment usage gpuweb/cts#1166.
-
Regroup memory_sync,buffer tests gpuweb/cts#1100.
-
Add presubmit check for long test names gpuweb/cts#1165.
-
Reduce the length of subresource test names gpuweb/cts#1164.
-
zero_init: add constant zero to prevent driver unrolling loops. gpuweb/cts#1163.
-
Add validation,resource_usages,texture,in_render_misc:* - Part II gpuweb/cts#1161.
-
Add validation,resource_usages,texture,in_render_misc:* - Part I gpuweb/cts#1157.
-
Test view format reinterpretation gpuweb/cts#1126.
-
Only 2D textures can be marked as render targets gpuweb/cts#1155.
-
operation,image_copy: Add support for depth16unorm gpuweb/cts#1154.
-
image_copy,mip_levels: reduce the mip level count. gpuweb/cts#1151.
-
depth_compare_func: Don't use 0.5 as depth16unorm rounds either way gpuweb/cts#1152.
-
Fix tests trying to copy aspects of depth-stencil textures. gpuweb/cts#1146.
-
depth_clip_clamp: only set stencilLoadOp and friends when needed. gpuweb/cts#1153.
-
Add validation,resource_usages,texture,in_render_common:* - Part V gpuweb/cts#1150.
-
val,buffer,create: Buffer sizes don't have to be multiple of 4. gpuweb/cts#1140.
-
image_copy: Ensure bufferOffset % 4 == 0 for depth-stencil. gpuweb/cts#1139.
-
Add validation,resource_usages,texture,in_render_common:* - Part IV gpuweb/cts#1132.
-
Fix infinite recursion in CopyExternalImageToTexture.spec.ts gpuweb/cts#1143.
-
Use globalThis instead of self. gpuweb/cts#1141.
-
val/copyT2T,format_compat: Don't assume compressed is always 4x4 gpuweb/cts#1134.
-
Fix DSAttachment load/store ops being specified when disallowed. gpuweb/cts#1136.
-
Fix test of BGL binding limits to correctly check limits gpuweb/cts#1137.
-
Fix a validation error caused by WriteTexture bytesPerRow: t.width gpuweb/cts#1133.
-
Add unicode tests for entryPoint matching. gpuweb/cts#1131.
-
writeBuffer: an unaligned bufferOffset is a validation error. gpuweb/cts#1129.
-
validation,vertex_state: Only check valid attribute @locations gpuweb/cts#1128.
-
Destroyed buffer submitting validation review and impl (#879) gpuweb/cts#1127.
-
-
7 pull requests closed.
-
draft for wgsl validation gpuweb/cts#755.
-
ShaderValidationTest: remove string overload gpuweb/cts#751.
-
shader/execution: Add tests for pipeline overrideable constants gpuweb/cts#794.
-
wgsl: test builtin operand types gpuweb/cts#724.
-
Defer subcase finalization until the end of the testcase gpuweb/cts#1170.
-
Update DeviceProvider to get multiple devices at once gpuweb/cts#1112.
-
Don't require offset % 4 == 0 for depth-stencil <-> buffer copies gpuweb/cts#1138.
-
-
46 pull requests received 183 new comments.
-
Move test device selection to an optional before clause gpuweb/cts#1333 (17 comments by @austinEng, @kainino0x)
-
Factor float numeric range generation into common location gpuweb/cts#1147 (13 comments by @ben-clayton, @zoddicus)
-
Don't block on subcases until the end of a testcase gpuweb/cts#1335 (7 comments by @austinEng, @kainino0x)
-
Implement OneULP and related ULP functions gpuweb/cts#1218 (7 comments by @ben-clayton, @dj2, @zoddicus)
-
Sparse color attachments validation tests gpuweb/cts#1149 (7 comments by @austinEng, @kainino0x, @shrekshao)
-
Update DevicePool to provide multiple devices gpuweb/cts#1142 (7 comments by @haoxli, @Jiawei-Shao, @kainino0x)
-
Fix DSAttachment load/store ops being specified when disallowed. gpuweb/cts#1136 (6 comments by @austinEng, @Kangz)
-
image_copy: Ensure bufferOffset % 4 == 0 for depth-stencil. gpuweb/cts#1139 (3 comments by @austinEng, @Kangz)
-
val/copyT2T,format_compat: Don't assume compressed is always 4x4 gpuweb/cts#1134 (3 comments by @Kangz, @lokokung)
-
Update pipeline output targets validation with different writeMask gpuweb/cts#1174 (2 comments by @austinEng, @shrekshao)
-
depth_clip_clamp: Fix validation error for stencil-less formats gpuweb/cts#1319 (2 comments by @Kangz, @toji)
-
Add a test with a Unicode character which would normalize gpuweb/cts#1194 (2 comments by @ben-clayton, @dj2)
-
shader/execution: Add tests for pipeline overrideable constants gpuweb/cts#794 (2 comments by @ben-clayton, @dj2)
-
Add validation,resource_usages,texture,in_render_misc:* - Part IV gpuweb/cts#1173 (2 comments by @jiangzhaoming)
-
val,buffer,create: Buffer sizes don't have to be multiple of 4. gpuweb/cts#1140 (2 comments by @kainino0x, @Kangz)
-
Don't require offset % 4 == 0 for depth-stencil <-> buffer copies gpuweb/cts#1138 (2 comments by @austinEng, @Kangz)
-
Destroyed buffer submitting validation review and impl (#879) gpuweb/cts#1127 (2 comments by @jiangzhaoming, @kainino0x)
-
Use subcases in subresources,buffer_usage_in_render_pass gpuweb/cts#1177 (1 comment by @jiangzhaoming)
-
val: depth/stencil readOnly states of render pass/bundle VS wrtie states of render pipeline gpuweb/cts#1171 (1 comment by @shrekshao)
-
Validation test for GPURenderPassDepthStencilAttachment usage gpuweb/cts#1166 (1 comment by @shrekshao)
-
Fix a bug in validation,createTexture,texture_usage:dimension=undef gpuweb/cts#1168 (1 comment by @Kangz)
TR publications
-
28 documents published.
-
WebGPU published 14 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 14 times as a Working Draft. History of WebGPU Shading Language
-
GitHub
gpuweb/gpuweb
-
31 issues opened.
-
Add scalar and vector constructors from abstract numeric types gpuweb/gpuweb#2719, by @dneto0.
-
WGSL parsing assumes context-aware tokenization gpuweb/gpuweb#2717, by @dneto0.
-
Disallow creating a single-layer 2d-array view of a 2d multisampled texture gpuweb/gpuweb#2715, by @austinEng.
-
Tracking issue for Bikeshed issues for the spec. gpuweb/gpuweb#2709, by @Kangz.
-
Requiring textures to be sized to a multiple of their block size is kind of unfortunate gpuweb/gpuweb#2706, by @litherum.
-
Add identity constructors for matrices gpuweb/gpuweb#2704, by @teoxoy.
-
Language around "physical size" of a texture is confused gpuweb/gpuweb#2698, by @litherum.
-
Task: Define the copies with 1d and 3d textures gpuweb/gpuweb#2690, by @litherum.
-
my2DArrayTexture.createView() doesn't create a full view gpuweb/gpuweb#2684, by @Kangz.
-
etc2-rgba8unorm and etc2-rgba8unorm-srgb do not exist in Metal gpuweb/gpuweb#2683, by @litherum.
-
Consider putting most of the writeTexture validation on the device timeline. gpuweb/gpuweb#2680, by @Kangz.
-
Make colon following switch case optional gpuweb/gpuweb#2678, by @amaiorano.
-
Proposal: Support DP4a as WGSL built-in functions gpuweb/gpuweb#2677, by @Jiawei-Shao.
-
Can't enable webgpu unsafe flag on Chrome Canary v102.0.4953.0 gpuweb/gpuweb#2674, by @eduardfossas.
-
Are Interleaved Vertex Buffers Preferred? gpuweb/gpuweb#2672, by @joeedh.
-
Clarify meaning of `dispatch()` args. gpuweb/gpuweb#2669, by @toji.
-
Why have literal suffixes at all? gpuweb/gpuweb#2664, by @kdashg.
-
What do the GPUBufferUsageFlags constants mean? gpuweb/gpuweb#2663, by @joeedh.
-
Are duplicate timestamp locations allowed on the same pass? gpuweb/gpuweb#2662, by @haoxli.
-
Can the identifier used in `enable` directive be keyword or reserved word? gpuweb/gpuweb#2649, by @jiangzhaoming.
-
Add rg11b10ufloat as valid render attachment format gpuweb/gpuweb#2648, by @pplux.
-
GPUBuffer.mapAsync is overly pessimistic gpuweb/gpuweb#2646, by @litherum.
-
Precision loss in timestamp query gpuweb/gpuweb#2642, by @haoxli.
-
Allow pipeline layouts to be partially automatic gpuweb/gpuweb#2637, by @kainino0x.
-
Make automatic bind group layouts explicit gpuweb/gpuweb#2636, by @kainino0x.
-
Concerns around error label and message gpuweb/gpuweb#2633, by @marcoscaceres.
-
wgsl: don't reserve certain words gpuweb/gpuweb#2632, by @dneto0.
-
Expose optional texture capabilities in WebGPU gpuweb/gpuweb#2630, by @Jiawei-Shao.
-
Remove serialization/transfer from the spec gpuweb/gpuweb#2629, by @kainino0x.
-
-
24 issues closed.
-
Language around "physical size" of a texture is confused gpuweb/gpuweb#2698.
-
Make colon following switch case optional gpuweb/gpuweb#2678.
-
Specify cube texture filtering as seamless gpuweb/gpuweb#2171.
-
Define edge case behaviour for f32 to f16 conversions gpuweb/gpuweb#1201.
-
W3C Reviews gpuweb/gpuweb#1321.
-
etc2-rgba8unorm and etc2-rgba8unorm-srgb do not exist in Metal gpuweb/gpuweb#2683.
-
Precision loss in timestamp query gpuweb/gpuweb#2642.
-
Should WGSL support Unicode identifiers? gpuweb/gpuweb#1160.
-
Should `clamp` be only defined for `low` <= `high`? gpuweb/gpuweb#2557.
-
Triage milestones for all issues gpuweb/gpuweb#2073.
-
Privacy Problem. gpuweb/gpuweb#2567.
-
Are Interleaved Vertex Buffers Preferred? gpuweb/gpuweb#2672.
-
Can't enable webgpu unsafe flag on Chrome Canary v102.0.4953.0 gpuweb/gpuweb#2674.
-
Should struct member declarations use commas instead of semicolons? gpuweb/gpuweb#2587.
-
Allow popErrorScope after device loss? gpuweb/gpuweb#2582.
-
Can the identifier used in `enable` directive be keyword or reserved word? gpuweb/gpuweb#2649.
-
Should unreachable-code be an error or warning? gpuweb/gpuweb#2378.
-
Consider removing GPUPipelineLayout gpuweb/gpuweb#2550.
-
consider adding: acosh, asinh, atanh gpuweb/gpuweb#1622.
-
67 issues received 269 new comments.
-
Should WGSL support Unicode identifiers? gpuweb/gpuweb#1160 (33 comments by @ben-clayton, @dneto0, @greggman, @kdashg, @litherum, @markusicu, @mehmetoguzderin)
-
Should `clamp` be only defined for `low` <= `high`? gpuweb/gpuweb#2557 (14 comments by @alan-baker, @dneto0, @kdashg, @kvark, @litherum, @pow2clk)
-
Concerns around error label and message gpuweb/gpuweb#2633 (11 comments by @kainino0x, @Kangz, @kdashg, @litherum, @marcoscaceres)
-
Add rg11b10ufloat as valid render attachment format gpuweb/gpuweb#2648 (11 comments by @kainino0x, @Kangz, @kdashg, @Kuranes, @litherum, @pplux)
-
Limit for the maximum buffer size gpuweb/gpuweb#1371 (11 comments by @kainino0x, @kdashg, @kvark, @litherum)
-
my2DArrayTexture.createView() doesn't create a full view gpuweb/gpuweb#2684 (10 comments by @kainino0x, @Kangz, @toji)
-
Streaming implementations and indirect draws/dispatches gpuweb/gpuweb#2189 (10 comments by @austinEng, @kainino0x, @kdashg, @litherum)
-
context-aware tokenization: Sometimes >> is better parsed as two copies of > (similar for ]] ) gpuweb/gpuweb#2092 (9 comments by @dneto0, @kdashg, @litherum)
-
Investigation: Generate mipmaps gpuweb/gpuweb#386 (9 comments by @davidar, @devshgraphicsprogramming, @Kangz)
-
mapSync on Workers - and possibly on the main thread gpuweb/gpuweb#2217 (8 comments by @greggman, @kainino0x, @kdashg, @litherum, @mrshannon, @toji)
-
Proposal of FP16 WGSL extension gpuweb/gpuweb#2512 (7 comments by @jiangzhaoming, @Kangz, @kdashg, @litherum)
-
Make error scope objects have same shape as DOMExceptions? gpuweb/gpuweb#1884 (6 comments by @juj, @kainino0x, @Kangz)
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520 (5 comments by @greggman, @kainino0x, @Kangz)
-
TextureView validation and defaults don't consider the aspect when checking/determining the view format gpuweb/gpuweb#2679 (5 comments by @austinEng, @Kangz)
-
GPUBuffer.mapAsync is overly pessimistic gpuweb/gpuweb#2646 (5 comments by @kainino0x, @Kangz, @litherum)
-
rgba8unorm canvas textures and IOSurface gpuweb/gpuweb#2535 (4 comments by @kainino0x, @Kangz, @litherum)
-
Consider removing GPUPipelineLayout gpuweb/gpuweb#2550 (4 comments by @austinEng, @kainino0x, @kdashg, @kvark)
-
Consider putting most of the writeTexture validation on the device timeline. gpuweb/gpuweb#2680 (3 comments by @kainino0x, @Kangz)
-
Should GPUDevice.destroy detach all buffer mapping ArrayBuffers? gpuweb/gpuweb#2545 (3 comments by @kainino0x)
-
Define edge case behaviour for f32 to f16 conversions gpuweb/gpuweb#1201 (3 comments by @dneto0, @kdashg, @litherum)
-
etc2-rgba8unorm and etc2-rgba8unorm-srgb do not exist in Metal gpuweb/gpuweb#2683 (3 comments by @Kangz, @litherum)
-
Should WGSL expose the concept of coherent buffers? gpuweb/gpuweb#1621 (3 comments by @devshgraphicsprogramming)
-
Validate that [[builtin fragDepth]] can't be written if depthWriteEnabled = false? gpuweb/gpuweb#1013 (3 comments by @Jiawei-Shao, @Kangz)
-
Does compositingAlphaMode affect using the canvas as an image source? gpuweb/gpuweb#1847 (3 comments by @kainino0x, @Kangz)
-
Cannot upload/download to UMA storage buffer without an unnecessary copy and unnecessary memory use gpuweb/gpuweb#2388 (3 comments by @Kangz, @kdashg)
-
Allow popErrorScope after device loss? gpuweb/gpuweb#2582 (3 comments by @kainino0x, @kdashg, @litherum)
-
Expose optional texture capabilities in WebGPU gpuweb/gpuweb#2630 (3 comments by @kainino0x, @kdashg)
-
Make colon following switch case optional gpuweb/gpuweb#2678 (2 comments by @kdashg, @mehmetoguzderin)
-
Should we constrain the location of user input-output stage variables WGSL gpuweb/gpuweb#1962 (2 comments by @kainino0x, @Kangz)
-
Requiring textures to be sized to a multiple of their block size is kind of unfortunate gpuweb/gpuweb#2706 (2 comments by @Jiawei-Shao, @Kangz)
-
Metal primitive topology required for the shader compilation hint gpuweb/gpuweb#2574 (2 comments by @cwfitzgerald, @Kangz)
-
Can't enable webgpu unsafe flag on Chrome Canary v102.0.4953.0 gpuweb/gpuweb#2674 (2 comments by @eduardfossas, @pixelspark)
-
Are duplicate timestamp locations allowed on the same pass? gpuweb/gpuweb#2662 (2 comments by @haoxli)
-
wgsl: Sketch design for arrays of buffer-bindings gpuweb/gpuweb#2105 (2 comments by @dneto0, @kdashg)
-
Fragment Built-in Output Value for Stencil [Feature Request] gpuweb/gpuweb#2553 (2 comments by @eurigilberto, @litherum)
-
Add scalar and vector constructors from abstract numeric types gpuweb/gpuweb#2719 (1 comment by @dneto0)
-
Proposal: Add a usage bit to allow creating a view with different format when creating a texture gpuweb/gpuweb#168 (1 comment by @Kangz)
-
Missing clarification on malicious use of GPU resources (e.g. mining) against intent of browser user gpuweb/gpuweb#2049 (1 comment by @aminick)
-
Exposing GPU hardware information to web developers in WebGPU gpuweb/gpuweb#2195 (1 comment by @toji)
-
Should struct member declarations use commas instead of semicolons? gpuweb/gpuweb#2587 (1 comment by @kdashg)
-
Can the identifier used in `enable` directive be keyword or reserved word? gpuweb/gpuweb#2649 (1 comment by @dneto0)
-
GPURenderBundleEncoderDescriptor needs maxCommandCount if it is to be implementable on MTLIndirectCommandBuffer gpuweb/gpuweb#2612 (1 comment by @austinEng)
-
-
53 pull requests opened.
-
Define WebCodecs interaction in an extension proposal gpuweb/gpuweb#2720, by @kainino0x.
-
Fix deprecation warnings in workload simulator gpuweb/gpuweb#2718, by @jdarpinian.
-
Require views on 2d multisampled textures to be 2d gpuweb/gpuweb#2716, by @austinEng.
-
Add identity constructors for matrix types gpuweb/gpuweb#2714, by @dneto0.
-
Remove shift tokens and add shift built-in functions gpuweb/gpuweb#2713, by @alan-baker.
-
Rename smoothStep to smoothstep gpuweb/gpuweb#2712, by @alan-baker.
-
Make colon in case optional gpuweb/gpuweb#2711, by @alan-baker.
-
Better specify the order of dynamicOffsets in setBindGroup gpuweb/gpuweb#2710, by @Kangz.
-
Specify that cubemaps are seamless gpuweb/gpuweb#2708, by @Kangz.
-
Add an "Abuse of capabilities" to malicious uses gpuweb/gpuweb#2707, by @Kangz.
-
Fix typo in uniformity analysis note gpuweb/gpuweb#2705, by @jrprice.
-
Align Python across repository gpuweb/gpuweb#2703, by @mehmetoguzderin.
-
Add the concept of an "aspect-specific format" gpuweb/gpuweb#2702, by @litherum.
-
Add rigorous definition of physical texture subresource size gpuweb/gpuweb#2701, by @litherum.
-
Removing links to closed issues gpuweb/gpuweb#2700, by @litherum.
-
Another small typo gpuweb/gpuweb#2699, by @litherum.
-
[editorial] Add definitions for uniformity tags gpuweb/gpuweb#2697, by @alan-baker.
-
FP16 Extension against spec gpuweb/gpuweb#2696, by @jiangzhaoming.
-
Fix typo: Code point should always be split gpuweb/gpuweb#2695, by @mehmetoguzderin.
-
Another typo gpuweb/gpuweb#2694, by @litherum.
-
Add texel block size column to format tables gpuweb/gpuweb#2693, by @litherum.
-
Another typo gpuweb/gpuweb#2692, by @litherum.
-
Rename dispatch -> dispatchWorkgroups gpuweb/gpuweb#2689, by @toji.
-
[editorial] Add an example for switch statements gpuweb/gpuweb#2688, by @alan-baker.
-
Clarify texture view validation and update defaults gpuweb/gpuweb#2687, by @austinEng.
-
wgsl: Fix a typo in Vector conversion type rules gpuweb/gpuweb#2686, by @jiangzhaoming.
-
Small typo gpuweb/gpuweb#2685, by @litherum.
-
WGSL: Explain parameterized type rules more explicitly. gpuweb/gpuweb#2682, by @jimblandy.
-
Floating clamp can be median-of-three gpuweb/gpuweb#2681, by @dneto0.
-
Error handling cleanup gpuweb/gpuweb#2675, by @litherum.
-
GPUSamplerDescriptor.maxAnisotropy gets clamped to a platform-specific maximum gpuweb/gpuweb#2670, by @litherum.
-
Add creation-time expressions gpuweb/gpuweb#2668, by @alan-baker.
-
Mipmap filtering might be extended separately from min/mag filtering in the future gpuweb/gpuweb#2667, by @litherum.
-
Commiting GPUCommandBuffers to the wrong GPUDevice is an error gpuweb/gpuweb#2666, by @litherum.
-
Add multi-planar-formats feature gpuweb/gpuweb#2665, by @jchen10.
-
Add GPUAdaperInfo to the spec gpuweb/gpuweb#2660, by @toji.
-
Make rg11b10ufloat RENDER_ATTACHMENT:yes and resolve:yes. gpuweb/gpuweb#2659, by @kdashg.
-
[editorial] Compress expression tables gpuweb/gpuweb#2658, by @alan-baker.
-
Make use of automatic pipeline layouts explicit gpuweb/gpuweb#2657, by @toji.
-
Uses commas to separate struct members intead of semicolons gpuweb/gpuweb#2656, by @alan-baker.
-
Typo in definition of finish() gpuweb/gpuweb#2655, by @litherum.
-
Use concise forms for entry point stage attributes. gpuweb/gpuweb#2652, by @jimblandy.
-
Resolve numeric literal types through inference. gpuweb/gpuweb#2651, by @jimblandy.
-
Name in enable directive can be a keyword or reserved word gpuweb/gpuweb#2650, by @dneto0.
-
FP16 Extension Document gpuweb/gpuweb#2647, by @jiangzhaoming.
-
Add a way of setting the initial label for the default queue gpuweb/gpuweb#2645, by @litherum.
-
make online should die on warnings gpuweb/gpuweb#2644, by @litherum.
-
GPUDevice.createBuffer() requires a valid GPUBufferDescriptor gpuweb/gpuweb#2643, by @litherum.
-
Reland: Only allow depth/stencil load/store ops when they have an effect gpuweb/gpuweb#2639, by @toji.
-
Add definitions for uniformity terms gpuweb/gpuweb#2638, by @alan-baker.
-
Allow unmasked adapter info fields to be requested individually. gpuweb/gpuweb#2635, by @toji.
-
wgsl: Fix typos of accuracy of exp and exp2 gpuweb/gpuweb#2634, by @jiangzhaoming.
-
-
48 pull requests merged.
-
Clarify texture view validation and update defaults gpuweb/gpuweb#2687.
-
Add rigorous definition of physical texture subresource size gpuweb/gpuweb#2701.
-
Add the concept of an "aspect-specific format" gpuweb/gpuweb#2702.
-
Fix deprecation warnings in workload simulator gpuweb/gpuweb#2718.
-
Make colon in case optional gpuweb/gpuweb#2711.
-
Remove shift tokens and add shift built-in functions gpuweb/gpuweb#2713.
-
Rename smoothStep to smoothstep gpuweb/gpuweb#2712.
-
Specify that cubemaps are seamless gpuweb/gpuweb#2708.
-
Literal unification, bottom-up gpuweb/gpuweb#2227.
-
editorial: discuss parsing ambiguity: type construction, function call gpuweb/gpuweb#2539.
-
Add an "Abuse of capabilities" to malicious uses gpuweb/gpuweb#2707.
-
Align Python across repository gpuweb/gpuweb#2703.
-
Fix typo in uniformity analysis note gpuweb/gpuweb#2705.
-
[editorial] Add definitions for uniformity tags gpuweb/gpuweb#2697.
-
Removing links to closed issues gpuweb/gpuweb#2700.
-
Another small typo gpuweb/gpuweb#2699.
-
Add texel block size column to format tables gpuweb/gpuweb#2693.
-
[editorial] Add an example for switch statements gpuweb/gpuweb#2688.
-
Fix typo: Code point should always be split gpuweb/gpuweb#2695.
-
Another typo gpuweb/gpuweb#2694.
-
Another typo gpuweb/gpuweb#2692.
-
Typo gpuweb/gpuweb#2691.
-
wgsl: Fix a typo in Vector conversion type rules gpuweb/gpuweb#2686.
-
Small typo gpuweb/gpuweb#2685.
-
Allow Unicode Identifiers gpuweb/gpuweb#2595.
-
Floating clamp can be median-of-three gpuweb/gpuweb#2681.
-
Error handling cleanup gpuweb/gpuweb#2675.
-
GPUSamplerDescriptor.maxAnisotropy gets clamped to a platform-specific maximum gpuweb/gpuweb#2670.
-
[Process] Add RequirementsForAdditionalFunctionality.md gpuweb/gpuweb#2424.
-
Add importExternalTexture examples with rAF/rVFC rendering gpuweb/gpuweb#2605.
-
Add a way of setting the initial label for the default queue gpuweb/gpuweb#2645.
-
Commiting GPUCommandBuffers to the wrong GPUDevice is an error gpuweb/gpuweb#2666.
-
make online should die on warnings gpuweb/gpuweb#2644.
-
[editorial] Compress expression tables gpuweb/gpuweb#2658.
-
Uses commas to separate struct members intead of semicolons gpuweb/gpuweb#2656.
-
Allow unmasked adapter info fields to be requested individually. gpuweb/gpuweb#2635.
-
Typo in definition of finish() gpuweb/gpuweb#2655.
-
GPUDevice.createBuffer() requires a valid GPUBufferDescriptor gpuweb/gpuweb#2643.
-
Name in enable directive can be a keyword or reserved word gpuweb/gpuweb#2650.
-
Allow statically unreachable code gpuweb/gpuweb#2622.
-
Add definitions for uniformity terms gpuweb/gpuweb#2638.
-
Validate query index overwrite in timestampWrites of render pass gpuweb/gpuweb#2627.
-
Reland: Only allow depth/stencil load/store ops when they have an effect gpuweb/gpuweb#2639.
-
wgsl: Fix typos of accuracy of exp and exp2 gpuweb/gpuweb#2634.
-
wgsl: Cleanup the internal memory layout of vector using SizeOf gpuweb/gpuweb#2626.
-
Add acosh, asinh, atanh builtin functions gpuweb/gpuweb#2581.
-
wgsl: Reserve words from common programming languages gpuweb/gpuweb#2617.
-
-
6 pull requests closed.
-
wgsl: Rename `smoothStep` to `smoothstep` gpuweb/gpuweb#2175.
-
createBindGroup: Require a superset of the layout's bindings gpuweb/gpuweb#1061.
-
Upload CI build output as artifact gpuweb/gpuweb#1184.
-
Experiment with using canvas.getContext() for configuration gpuweb/gpuweb#2593.
-
Add multi-planar-formats feature gpuweb/gpuweb#2665.
-
Remove createPipelineLayout, take layout "requirements" by value gpuweb/gpuweb#2572.
-
-
49 pull requests received 245 new comments.
-
Add creation-time expressions gpuweb/gpuweb#2668 (28 comments by @alan-baker, @dneto0, @jrprice, @kdashg)
-
Literal unification, bottom-up gpuweb/gpuweb#2227 (22 comments by @alan-baker, @dneto0, @jimblandy, @kdashg)
-
Add rigorous definition of physical texture subresource size gpuweb/gpuweb#2701 (13 comments by @kainino0x, @Kangz, @litherum)
-
Rely on canvas size instead of allowing size to be configured gpuweb/gpuweb#2416 (13 comments by @annevk, @kainino0x, @toji)
-
FP16 Extension against spec gpuweb/gpuweb#2696 (11 comments by @alan-baker, @jiangzhaoming, @kainino0x, @Kangz)
-
Resolve numeric literal types through inference. gpuweb/gpuweb#2651 (11 comments by @alan-baker, @dneto0, @jimblandy, @kdashg, @RobinMorisset)
-
Allow unmasked adapter info fields to be requested individually. gpuweb/gpuweb#2635 (7 comments by @kainino0x, @kdashg, @toji)
-
Add the concept of an "aspect-specific format" gpuweb/gpuweb#2702 (6 comments by @austinEng, @Kangz, @litherum)
-
GPUSamplerDescriptor.maxAnisotropy gets clamped to a platform-specific maximum gpuweb/gpuweb#2670 (6 comments by @kainino0x, @litherum)
-
[Process] Add RequirementsForAdditionalFunctionality.md gpuweb/gpuweb#2424 (6 comments by @kainino0x, @kdashg, @litherum)
-
Add a way of setting the initial label for the default queue gpuweb/gpuweb#2645 (6 comments by @kainino0x, @litherum, @toji)
-
Add proposals/ directory and pipeline statistics extension doc gpuweb/gpuweb#2301 (5 comments by @kainino0x, @kdashg)
-
Clarify texture view validation and update defaults gpuweb/gpuweb#2687 (4 comments by @austinEng, @kainino0x, @toji)
-
Experiment with using canvas.getContext() for configuration gpuweb/gpuweb#2593 (4 comments by @annevk, @kainino0x, @toji)
-
Allow statically unreachable code gpuweb/gpuweb#2622 (4 comments by @alan-baker, @dneto0, @kdashg, @RobinMorisset)
-
Validate query index overwrite in timestampWrites of render pass gpuweb/gpuweb#2627 (4 comments by @haoxli, @kainino0x)
-
WGSL: Explain parameterized type rules more explicitly. gpuweb/gpuweb#2682 (3 comments by @alan-baker, @dneto0)
-
Make rg11b10ufloat RENDER_ATTACHMENT:yes and resolve:yes. gpuweb/gpuweb#2659 (3 comments by @kainino0x, @kdashg, @mrshannon)
-
Correct the definition of fragment-stage position builtin gpuweb/gpuweb#2625 (3 comments by @dneto0, @kdashg, @surma)
-
wgsl: Reserve words from common programming languages gpuweb/gpuweb#2617 (3 comments by @dneto0, @jimblandy, @kdashg)
-
Require views on 2d multisampled textures to be 2d gpuweb/gpuweb#2716 (2 comments by @austinEng, @kainino0x)
-
editorial: discuss parsing ambiguity: type construction, function call gpuweb/gpuweb#2539 (2 comments by @dneto0, @kdashg)
-
Mipmap filtering might be extended separately from min/mag filtering in the future gpuweb/gpuweb#2667 (2 comments by @kainino0x, @Kangz)
-
Make use of automatic pipeline layouts explicit gpuweb/gpuweb#2657 (2 comments by @kainino0x, @toji)
-
Allow e.g. `if (cond) break/continue/return/kill;` without brackets. gpuweb/gpuweb#705 (1 comment by @Kangz)
-
Commiting GPUCommandBuffers to the wrong GPUDevice is an error gpuweb/gpuweb#2666 (1 comment by @toji)
-
Name in enable directive can be a keyword or reserved word gpuweb/gpuweb#2650 (1 comment by @kdashg)
-
Add importExternalTexture examples with rAF/rVFC rendering gpuweb/gpuweb#2605 (1 comment by @kainino0x)
-
Remove createPipelineLayout, take layout "requirements" by value gpuweb/gpuweb#2572 (1 comment by @toji)
-
gpuweb/cts
-
19 issues opened.
-
wgsl: Add tests for builtin `exp` gpuweb/cts#1125, by @zoddicus.
-
helpers: use texture_ok helpers in more tests gpuweb/cts#1121, by @kainino0x.
-
f32 arithmetic tests need to account for subnormals and flushing behaviours gpuweb/cts#1118, by @zoddicus.
-
reftest: canvas scaling gpuweb/cts#1116, by @kainino0x.
-
reftest: canvas sizing with ResizeObserver and checkerboard pattern gpuweb/cts#1115, by @kainino0x.
-
op: error_scope gpuweb/cts#1111, by @lokokung.
-
Replace `Array.concat` with `Array.push` in builtin tests gpuweb/cts#1103, by @zoddicus.
-
wgsl: Update `clamp.spec.ts` to match spec changes gpuweb/cts#1101, by @zoddicus.
-
Add canvas tests for RGBA16Float values with regard to SDR luminance gpuweb/cts#1093, by @jchen10.
-
`npm run wpt` does not populate `out-wpt/common` fully gpuweb/cts#1092, by @litherum.
-
Show some progress when loading the CTS tests gpuweb/cts#1079, by @greggman.
-
val: createTexture: sampleCount is invalid with 1d/3d textures gpuweb/cts#1078, by @greggman.
-
op: explicitly test WebGPU is available on DedicatedWorkers gpuweb/cts#1066, by @greggman.
-
op: requestAdapter gpuweb/cts#1061, by @greggman.
-
ESLint requires LF - change git settings to always checkout as LF gpuweb/cts#1047, by @rconde01.
-
web_platform: reference tests imported to WPT need to be fuzzy gpuweb/cts#1045, by @Juanmihd.
-
wgsl: Names in shader builtin tests duplicate information gpuweb/cts#1039, by @zoddicus.
-
op: buffers,map_ArrayBuffer gpuweb/cts#1033, by @kainino0x.
-
Move pipeline statistics tests to another suite? gpuweb/cts#1031, by @kainino0x.
-
-
11 issues closed.
-
val: required limits must be a power of 2 gpuweb/cts#939.
-
Consolidate helpers for texture content checking gpuweb/cts#881.
-
Replace `Array.concat` with `Array.push` in builtin tests gpuweb/cts#1103.
-
wgsl: Update `clamp.spec.ts` to match spec changes gpuweb/cts#1101.
-
op: memory_sync,buffer,rw_and_wr gpuweb/cts#887.
-
val: encoding,render_bundle gpuweb/cts#1011.
-
wgsl: Names in shader builtin tests duplicate information gpuweb/cts#1039.
-
ESLint requires LF - change git settings to always checkout as LF gpuweb/cts#1047.
-
op: copyTextureToTexture between srgb and non-srgb gpuweb/cts#935.
-
val: beginRenderPass resolveTarget format must support resolve gpuweb/cts#940.
-
-
8 issues received 10 new comments.
-
Show some progress when loading the CTS tests gpuweb/cts#1079 (2 comments by @austinEng, @kainino0x)
-
web_platform: reference tests imported to WPT need to be fuzzy gpuweb/cts#1045 (2 comments by @Juanmihd, @kainino0x)
-
val: capability_checks,features errors should be exceptions gpuweb/cts#919 (1 comment by @kainino0x)
-
-
84 pull requests opened.
-
Test view format reinterpretation gpuweb/cts#1126, by @austinEng.
-
Better check for presence of the Web platform gpuweb/cts#1124, by @Kangz.
-
Improvements for running in node. gpuweb/cts#1123, by @Kangz.
-
Make an info table for limits. gpuweb/cts#1122, by @Kangz.
-
Update deprecated WGSL struct syntax (";" -> ",") gpuweb/cts#1120, by @toji.
-
Add handling subnormals to f32 arithmetic tests gpuweb/cts#1119, by @zoddicus.
-
Only request STORAGE_BINDING for store cases gpuweb/cts#1117, by @jchen10.
-
Fix RenderBundle read-only state mismatch test gpuweb/cts#1114, by @toji.
-
Remove scaling for canvas_composite_alpha gpuweb/cts#1113, by @jchen10.
-
Update DeviceProvider to get multiple devices at once gpuweb/cts#1112, by @haoxli.
-
Fix numeric array sort in WGSL clamp tests gpuweb/cts#1110, by @jrprice.
-
Fix half-pixel for canvas_composite_alpha gpuweb/cts#1109, by @jchen10.
-
Remove scaling for canvas_complex gpuweb/cts#1108, by @jchen10.
-
Remove test leading to FXC compile failure gpuweb/cts#1107, by @amaiorano.
-
Implements remaining device.destroy validation tests. gpuweb/cts#1106, by @lokokung.
-
Add remaining device.destroy tests. gpuweb/cts#1105, by @lokokung.
-
Replace usages of `concat` with `push` gpuweb/cts#1104, by @zoddicus.
-
Update `clamp` tests to match spec gpuweb/cts#1102, by @zoddicus.
-
Regroup memory_sync,buffer tests gpuweb/cts#1100, by @shrekshao.
-
wgsl: Convert last remaining [[]] attribute. gpuweb/cts#1099, by @Kangz.
-
Add validation,resource_usages,texture,in_render_common:* - Part III gpuweb/cts#1098, by @Jiawei-Shao.
-
Fix floating point arithmetic tests so they actually run gpuweb/cts#1097, by @zoddicus.
-
npm update, upgrade lockfileVersion, remove git conflict marker from lockfile gpuweb/cts#1096, by @kainino0x.
-
Implement tests for floating point logical expressions gpuweb/cts#1095, by @zoddicus.
-
Implement api,validation,state,device_mismatched: Part II gpuweb/cts#1094, by @haoxli.
-
memory_sync,buffer,rw_and_wr gpuweb/cts#1091, by @shrekshao.
-
Add checkbox to auto-close subtrees when they pass gpuweb/cts#1090, by @kainino0x.
-
color_space_conversion: use exact rational matrix formulations gpuweb/cts#1089, by @kainino0x.
-
Fix readSinglePixelFrom2DTexture helper gpuweb/cts#1088, by @kainino0x.
-
wgsl: Remove usage of @stride attribute gpuweb/cts#1087, by @jrprice.
-
Fixing issues with fuzziness gpuweb/cts#1086, by @Juanmihd.
-
Fixing fuzziness gpuweb/cts#1085, by @Juanmihd.
-
Fix Spelling Errors gpuweb/cts#1084, by @greggman.
-
Add evaluation order tests gpuweb/cts#1083, by @amaiorano.
-
Skip new depth/stencil tests that require optional formats gpuweb/cts#1082, by @austinEng.
-
Add canvas STORAGE_BINDING cases for RGBA8Unorm and RGBA16Float gpuweb/cts#1081, by @jchen10.
-
Show some loading progress gpuweb/cts#1080, by @greggman.
-
Bugfix for getTextureCopyLayout gpuweb/cts#1077, by @kainino0x.
-
wgsl: Add validation tests for tokenization gpuweb/cts#1076, by @jrprice.
-
Implement Basic Worker WebGPU Test gpuweb/cts#1075, by @greggman.
-
Fix Spelling Errors gpuweb/cts#1074, by @greggman.
-
wgsl: Implement tests for floating point arithmetic expressions gpuweb/cts#1073, by @zoddicus.
-
Reduce instances of `// prettier-ignore` to improve readability gpuweb/cts#1072, by @zoddicus.
-
Implement `reverseBits` tests gpuweb/cts#1071, by @zoddicus.
-
Add validation,resource_usages,texture,in_render_common:* - Part II gpuweb/cts#1070, by @Jiawei-Shao.
-
Change endPass to end in in_render_common.spec.ts gpuweb/cts#1069, by @Jiawei-Shao.
-
Assorted helpers used for texture checking gpuweb/cts#1068, by @kainino0x.
-
simplify physicalMipSize and virtualMipSize gpuweb/cts#1067, by @kainino0x.
-
Implement `countOneBits` tests gpuweb/cts#1065, by @zoddicus.
-
Bitwise ops gpuweb/cts#1064, by @ben-clayton.
-
Remove test stubs for `isInfinite` and `isNormal` gpuweb/cts#1063, by @zoddicus.
-
Reduce duplication of terms in builtin test strings gpuweb/cts#1062, by @zoddicus.
-
Fix limits of `log()` and `log2()` gpuweb/cts#1060, by @ben-clayton.
-
webgpu/shader/execution: Shuffle harness code gpuweb/cts#1059, by @ben-clayton.
-
Implement CTS requestAdapter test gpuweb/cts#1058, by @greggman.
-
Add validation tests for render bundle creation and usage gpuweb/cts#1057, by @toji.
-
Enable multi-planar-formats ImportExternalTexture tests gpuweb/cts#1056, by @jchen10.
-
Helper for robust texture content checking gpuweb/cts#1055, by @kainino0x.
-
replaced another expired origin trial token gpuweb/cts#1054, by @kainino0x.
-
Attempt to wait for queue idle on release of the device to the pool gpuweb/cts#1053, by @austinEng.
-
Increase fuzziness gpuweb/cts#1052, by @Juanmihd.
-
Don't set depth/stencil load/store ops if no respective aspect. Pt 2 gpuweb/cts#1051, by @austinEng.
-
Roll TypeScript gpuweb/cts#1050, by @kainino0x.
-
Don't set depth/stencil load/store ops if the aspect doesn't exist gpuweb/cts#1049, by @austinEng.
-
update git attributes to check out text files with lf eol regardless … gpuweb/cts#1048, by @rconde01.
-
Fuzzy canvas reftests gpuweb/cts#1046, by @Juanmihd.
-
webgpu: Remove use of `isNan` and `isInf` gpuweb/cts#1044, by @ben-clayton.
-
CTS: Color space conversion operation test for CopyToTexture gpuweb/cts#1043, by @shaoboyan.
-
Update a bunch of dependencies gpuweb/cts#1042, by @greggman.
-
Fix api,operation,programmable,state_tracking - again gpuweb/cts#1041, by @austinEng.
-
Add validation,resource_usages,texture,in_render_common:* - Part I gpuweb/cts#1040, by @Jiawei-Shao.
-
webgpu: Migrate more `endPass()` -> `end()` gpuweb/cts#1038, by @ben-clayton.
-
Match layout and shader in command_buffer,programmable,state_tracking gpuweb/cts#1037, by @austinEng.
-
Add `log2` tests gpuweb/cts#1036, by @zoddicus.
-
Add `log` tests gpuweb/cts#1035, by @zoddicus.
-
map_ArrayBuffer.spec.ts: initialize contents before mapping gpuweb/cts#1034, by @austinEng.
-
Add `ldexp` tests gpuweb/cts#1032, by @zoddicus.
-
Update to latest WebGPU CTS API gpuweb/cts#1030, by @toji.
-
Remove deprecated copyImageBitmapToTexture from comments and readme gpuweb/cts#1029, by @takahirox.
-
Add variants of canvas_composite_alpha reftests that use a copy gpuweb/cts#1028, by @austinEng.
-
buffers,map_ArrayBuffer:* fake transfer of mapped array buffers gpuweb/cts#1027, by @austinEng.
-
Implement `sin` tests gpuweb/cts#1025, by @zoddicus.
-
Ensure lost device messages output the error message and reason. gpuweb/cts#1024, by @toji.
-
-
80 pull requests merged.
-
Better check for presence of the Web platform gpuweb/cts#1124.
-
Improvements for running in node. gpuweb/cts#1123.
-
Make an info table for limits. gpuweb/cts#1122.
-
Only request STORAGE_BINDING for store cases gpuweb/cts#1117.
-
Update deprecated WGSL struct syntax (";" -> ",") gpuweb/cts#1120.
-
Add handling subnormals to f32 arithmetic tests gpuweb/cts#1119.
-
Remove scaling for canvas_composite_alpha gpuweb/cts#1113.
-
Helper for robust texture content checking gpuweb/cts#1055.
-
Fix RenderBundle read-only state mismatch test gpuweb/cts#1114.
-
Fix half-pixel for canvas_composite_alpha gpuweb/cts#1109.
-
Remove scaling for canvas_complex gpuweb/cts#1108.
-
Replace usages of `concat` with `push` gpuweb/cts#1104.
-
Implements remaining device.destroy validation tests. gpuweb/cts#1106.
-
Fix numeric array sort in WGSL clamp tests gpuweb/cts#1110.
-
Remove test leading to FXC compile failure gpuweb/cts#1107.
-
Update `clamp` tests to match spec gpuweb/cts#1102.
-
Add evaluation order tests gpuweb/cts#1083.
-
Implement api,validation,state,device_mismatched: Part II gpuweb/cts#1094.
-
Add validation,resource_usages,texture,in_render_common:* - Part III gpuweb/cts#1098.
-
Add canvas STORAGE_BINDING cases for RGBA8Unorm and RGBA16Float gpuweb/cts#1081.
-
Adds destroy validation tests for some APIs. gpuweb/cts#977.
-
Implement tests for floating point logical expressions gpuweb/cts#1095.
-
Fix floating point arithmetic tests so they actually run gpuweb/cts#1097.
-
wgsl: Convert last remaining [[]] attribute. gpuweb/cts#1099.
-
memory_sync,buffer,rw_and_wr gpuweb/cts#1091.
-
color_space_conversion: use exact rational matrix formulations gpuweb/cts#1089.
-
Show some loading progress gpuweb/cts#1080.
-
wgsl: Add validation tests for tokenization gpuweb/cts#1076.
-
wgsl: Remove usage of @stride attribute gpuweb/cts#1087.
-
Add checkbox to auto-close subtrees when they pass gpuweb/cts#1090.
-
Fix Spelling Errors gpuweb/cts#1084.
-
Fix readSinglePixelFrom2DTexture helper gpuweb/cts#1088.
-
Skip new depth/stencil tests that require optional formats gpuweb/cts#1082.
-
Fixing issues with fuzziness gpuweb/cts#1086.
-
Bugfix for getTextureCopyLayout gpuweb/cts#1077.
-
Implement Basic Worker WebGPU Test gpuweb/cts#1075.
-
wgsl: Implement tests for floating point arithmetic expressions gpuweb/cts#1073.
-
Add validation,resource_usages,texture,in_render_common:* - Part II gpuweb/cts#1070.
-
Fix Spelling Errors gpuweb/cts#1074.
-
Assorted helpers used for texture checking gpuweb/cts#1068.
-
Add validation tests for render bundle creation and usage gpuweb/cts#1057.
-
Reduce instances of `// prettier-ignore` to improve readability gpuweb/cts#1072.
-
Implement `reverseBits` tests gpuweb/cts#1071.
-
Change endPass to end in in_render_common.spec.ts gpuweb/cts#1069.
-
Reduce duplication of terms in builtin test strings gpuweb/cts#1062.
-
Bitwise ops gpuweb/cts#1064.
-
Implement `countOneBits` tests gpuweb/cts#1065.
-
Remove test stubs for `isInfinite` and `isNormal` gpuweb/cts#1063.
-
CTS: Color space conversion operation test for CopyToTexture gpuweb/cts#1043.
-
Fix limits of `log()` and `log2()` gpuweb/cts#1060.
-
webgpu/shader/execution: Shuffle harness code gpuweb/cts#1059.
-
Increase fuzziness gpuweb/cts#1052.
-
replaced another expired origin trial token gpuweb/cts#1054.
-
Don't set depth/stencil load/store ops if no respective aspect. Pt 2 gpuweb/cts#1051.
-
Attempt to wait for queue idle on release of the device to the pool gpuweb/cts#1053.
-
Roll TypeScript gpuweb/cts#1050.
-
Don't set depth/stencil load/store ops if the aspect doesn't exist gpuweb/cts#1049.
-
Add validation,resource_usages,texture,in_render_common:* - Part I gpuweb/cts#1040.
-
Fuzzy canvas reftests gpuweb/cts#1046.
-
webgpu: Remove use of `isNan` and `isInf` gpuweb/cts#1044.
-
Update a bunch of dependencies gpuweb/cts#1042.
-
Fix api,operation,programmable,state_tracking - again gpuweb/cts#1041.
-
Add variants of canvas_composite_alpha reftests that use a copy gpuweb/cts#1028.
-
map_ArrayBuffer.spec.ts: initialize contents before mapping gpuweb/cts#1034.
-
Add `ldexp` tests gpuweb/cts#1032.
-
Add `log` tests gpuweb/cts#1035.
-
Add `log2` tests gpuweb/cts#1036.
-
webgpu: Migrate more `endPass()` -> `end()` gpuweb/cts#1038.
-
Implement `sin` tests gpuweb/cts#1025.
-
buffers,map_ArrayBuffer:* fake transfer of mapped array buffers gpuweb/cts#1027.
-
Implement `atan2` test gpuweb/cts#1023.
-
Update to latest WebGPU CTS API gpuweb/cts#1030.
-
Remove deprecated copyImageBitmapToTexture from comments and readme gpuweb/cts#1029.
-
Add additional memory model tests and refactor setup code gpuweb/cts#1022.
-
Ensure lost device messages output the error message and reason. gpuweb/cts#1024.
-
Add validation tests on resolve target formats gpuweb/cts#1021.
-
-
4 pull requests closed.
-
Add remaining device.destroy tests. gpuweb/cts#1105.
-
Fixing fuzziness gpuweb/cts#1085.
-
Enable multi-planar-formats ImportExternalTexture tests gpuweb/cts#1056.
-
Update to latest WebGPU CTS API gpuweb/cts#984.
-
-
52 pull requests received 365 new comments.
-
CTS: Color space conversion operation test for CopyToTexture gpuweb/cts#1043 (57 comments by @austinEng, @kainino0x, @shaoboyan)
-
Helper for robust texture content checking gpuweb/cts#1055 (40 comments by @austinEng, @kainino0x, @shaoboyan)
-
Assorted helpers used for texture checking gpuweb/cts#1068 (25 comments by @austinEng, @kainino0x, @shaoboyan)
-
Add validation tests for render bundle creation and usage gpuweb/cts#1057 (19 comments by @austinEng, @kainino0x, @toji)
-
Replace usages of `concat` with `push` gpuweb/cts#1104 (12 comments by @austinEng, @ben-clayton, @dneto0, @zoddicus)
-
Implement api,validation,state,device_mismatched: Part II gpuweb/cts#1094 (12 comments by @haoxli, @kainino0x)
-
wgsl: Implement tests for floating point arithmetic expressions gpuweb/cts#1073 (9 comments by @ben-clayton, @zoddicus)
-
Add canvas STORAGE_BINDING cases for RGBA8Unorm and RGBA16Float gpuweb/cts#1081 (8 comments by @jchen10, @kainino0x)
-
Complete non-single-pass webgpu:api,operation,memory_sync,texture,same_subresource:* gpuweb/cts#1026 (7 comments by @austinEng, @kainino0x)
-
Add validation,resource_usages,texture,in_render_common:* - Part I gpuweb/cts#1040 (7 comments by @austinEng, @Jiawei-Shao)
-
Implements remaining device.destroy validation tests. gpuweb/cts#1106 (6 comments by @austinEng, @lokokung)
-
Add validation,resource_usages,texture,in_render_common:* - Part II gpuweb/cts#1070 (6 comments by @Jiawei-Shao, @kainino0x)
-
color_space_conversion: use exact rational matrix formulations gpuweb/cts#1089 (5 comments by @kainino0x, @shaoboyan)
-
wgsl: Add validation tests for tokenization gpuweb/cts#1076 (5 comments by @jrprice, @kainino0x, @zoddicus)
-
Reduce duplication of terms in builtin test strings gpuweb/cts#1062 (5 comments by @ben-clayton, @zoddicus)
-
buffers,map_ArrayBuffer:* fake transfer of mapped array buffers gpuweb/cts#1027 (5 comments by @austinEng, @kainino0x)
-
Implement tests for floating point logical expressions gpuweb/cts#1095 (4 comments by @ben-clayton, @zoddicus)
-
Attempt to wait for queue idle on release of the device to the pool gpuweb/cts#1053 (4 comments by @austinEng, @kainino0x)
-
Match layout and shader in command_buffer,programmable,state_tracking gpuweb/cts#1037 (4 comments by @austinEng, @kainino0x)
-
Add variants of canvas_composite_alpha reftests that use a copy gpuweb/cts#1028 (3 comments by @austinEng, @kainino0x)
-
Add validation,resource_usages,texture,in_render_common:* - Part III gpuweb/cts#1098 (2 comments by @Kangz)
-
Ensure lost device messages output the error message and reason. gpuweb/cts#1024 (2 comments by @austinEng, @toji)
-
npm update, upgrade lockfileVersion, remove git conflict marker from lockfile gpuweb/cts#1096 (1 comment by @greggman)
-
TR publications
-
34 documents published.
-
WebGPU published 17 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 17 times as a Working Draft. History of WebGPU Shading Language
-
Participants
-
1 participant joined.
-
Xiaojian Wang (Alibaba Group) joined.
-
-
1 organization joined.
-
Alibaba Group joined.
-
GitHub
gpuweb/gpuweb
-
30 issues opened.
-
change endPass() to end() in IDL gpuweb/gpuweb#2620, by @dneto0.
-
Rules for shadowing built-in functions are unclear gpuweb/gpuweb#2619, by @jimblandy.
-
WGSL vs WebGPU spec for override constants gpuweb/gpuweb#2613, by @brendan-duncan.
-
need schema for features still under development vs. finalized gpuweb/gpuweb#2608, by @dneto0.
-
Consider reserving keywords for a module system gpuweb/gpuweb#2607, by @dneto0.
-
WriteBuffer performance issue? gpuweb/gpuweb#2601, by @lslzl3000.
-
Make GPUDevice.destroy() a queued operation gpuweb/gpuweb#2599, by @kainino0x.
-
Consider reserving the `std` or `wgsl` identifiers gpuweb/gpuweb#2591, by @Kangz.
-
Concise attributes gpuweb/gpuweb#2589, by @alan-baker.
-
Integer signedness usability gpuweb/gpuweb#2588, by @alan-baker.
-
Should struct member declarations use commas instead of semicolons? gpuweb/gpuweb#2587, by @alan-baker.
-
Workgroup Broadcast gpuweb/gpuweb#2586, by @alan-baker.
-
Static assert gpuweb/gpuweb#2584, by @alan-baker.
-
Allow popErrorScope after device loss? gpuweb/gpuweb#2582, by @kainino0x.
-
Add a MatrixInverse builtin? gpuweb/gpuweb#2580, by @kdashg.
-
Reconsider adding `while` to the language gpuweb/gpuweb#2578, by @RobinMorisset.
-
Proposal: Make `;` optional gpuweb/gpuweb#2577, by @ben-clayton.
-
Proposal: Make `(` `)` optional for `if`, `switch` and `for` gpuweb/gpuweb#2575, by @ben-clayton.
-
Metal primitive topology required for the shader compilation hint gpuweb/gpuweb#2574, by @kvark.
-
GPUTextureFormatDescriptor.viewFormats performance hint is nonhelpful gpuweb/gpuweb#2573, by @juj.
-
Binary WGSL format in createShaderModule() WebAPI gpuweb/gpuweb#2571, by @munrocket.
-
Integer signedness consistency gpuweb/gpuweb#2570, by @kvark.
-
"validating linear texture data" should depend on the aspect for depth-stencil formats gpuweb/gpuweb#2569, by @Kangz.
-
Privacy Problem. gpuweb/gpuweb#2567, by @Rootkage.
-
Query texture format caps from the device gpuweb/gpuweb#2565, by @brendan-duncan.
-
Dynamic indexing in Uniform address space gpuweb/gpuweb#2559, by @kvark.
-
Add a note that copyT2T might not preserve the exact bit values gpuweb/gpuweb#2558, by @Kangz.
-
Should `clamp` be only defined for `low` <= `high`? gpuweb/gpuweb#2557, by @zoddicus.
-
Consider renaming endPass() to end() gpuweb/gpuweb#2555, by @Kangz.
-
-
63 issues closed.
-
Rules about "break" inside "continuing" are strange gpuweb/gpuweb#1867.
-
Rules for shadowing built-in functions are unclear gpuweb/gpuweb#2619.
-
change endPass() to end() in IDL gpuweb/gpuweb#2620.
-
need schema for features still under development vs. finalized gpuweb/gpuweb#2608.
-
WGSL vs WebGPU spec for override constants gpuweb/gpuweb#2613.
-
Proposal: turn "const" at global level into a real storage class gpuweb/gpuweb#1394.
-
Consider reserving the `std` or `wgsl` identifiers gpuweb/gpuweb#2591.
-
Dynamic indexing in Uniform address space gpuweb/gpuweb#2559.
-
Consider not requiring braces in some situations gpuweb/gpuweb#2211.
-
WriteBuffer performance issue? gpuweb/gpuweb#2601.
-
Add a MatrixInverse builtin? gpuweb/gpuweb#2580.
-
Separate the privacy and security consideration sections of the spec. gpuweb/gpuweb#2547.
-
Sparse colorAttachments in a render pass gpuweb/gpuweb#1250.
-
Feature request: ignore shader writes to color attachments gpuweb/gpuweb#2060.
-
Reconsider adding `while` to the language gpuweb/gpuweb#2578.
-
Proposal: Make `(` `)` optional for `if`, `switch` and `for` gpuweb/gpuweb#2575.
-
Investigate remaining GLSL.std.450 methods for inclusion. gpuweb/gpuweb#1087.
-
WGSL type rules need to say something about aliases, and about attributes gpuweb/gpuweb#1387.
-
WGSL should require storage buffers to be a host-shareable storable type. gpuweb/gpuweb#877.
-
Task: Define pointers gpuweb/gpuweb#672.
-
Task: Finish typing rules gpuweb/gpuweb#664.
-
MSL question: can mem_flags be combined in threadgroup_barrier? gpuweb/gpuweb#1547.
-
builtin attributes on struct members gpuweb/gpuweb#1981.
-
Task: Invent a standard library gpuweb/gpuweb#667.
-
Task: Variable scoping is under-defined gpuweb/gpuweb#666.
-
Task: Validation rules gpuweb/gpuweb#665.
-
Task: Define order of execution gpuweb/gpuweb#663.
-
Task: Uniformity gpuweb/gpuweb#669.
-
Task: Numerical precision gpuweb/gpuweb#674.
-
[WGSL] Validate the grammar is unambiguous gpuweb/gpuweb#1063.
-
Mechanism for opting into shading language extensions gpuweb/gpuweb#1404.
-
[wgsl] Define what bijective to SPIR-V means gpuweb/gpuweb#582.
-
[proposal] Changes to overridable constants gpuweb/gpuweb#2238.
-
Develop a visual language for the syntactic rules in WGSL specification gpuweb/gpuweb#2114.
-
further restrict where a control-flow interruption can be placed gpuweb/gpuweb#2093.
-
Out of bounds behaviour and atomics gpuweb/gpuweb#1868.
-
wgsl: limiting undefined behaviour; subdividing "dynamic error" gpuweb/gpuweb#1844.
-
[wgsl] continuing block gpuweb/gpuweb#579.
-
[wgsl] Define the interface of an entry point in a WGSL program gpuweb/gpuweb#774.
-
Consider whether to allow trailing commas gpuweb/gpuweb#1243.
-
Proposal: Make `;` optional gpuweb/gpuweb#2577.
-
SPIR-V doesn't grant reconvergence for multiple returns gpuweb/gpuweb#1770.
-
[wgsl] Consider a preprocessor gpuweb/gpuweb#568.
-
Consider renaming endPass() to end() gpuweb/gpuweb#2555.
-
Validation of whole shader modules gpuweb/gpuweb#974.
-
editorial: I'd like to write WGSL instead of [SHORTNAME] gpuweb/gpuweb#1589.
-
Float literal scientific notation? gpuweb/gpuweb#1432.
-
investigation: texture load channel values for channels not in the texel format gpuweb/gpuweb#1115.
-
Device.getShaderPrecisionFormat(shaderType, precisionType) gpuweb/gpuweb#1048.
-
Shader portability (and performance portability). gpuweb/gpuweb#895.
-
WorkGroup memory size and barriers? gpuweb/gpuweb#863.
-
Define "human writable" as "can be written by humans" gpuweb/gpuweb#681.
-
[wgsl] Numeric sizes gpuweb/gpuweb#575.
-
Flush-to-zero denormal behavior gpuweb/gpuweb#349.
-
Document behaviour when uniformity requirement can't be proven gpuweb/gpuweb#340.
-
behaviour of out-of-bounds bitfield extract and insert gpuweb/gpuweb#288.
-
shader programming model: permitted memory orders on control barriers gpuweb/gpuweb#232.
-
security model, focused on shaders gpuweb/gpuweb#39.
-
-
95 issues received 326 new comments.
-
Should WGSL support Unicode identifiers? gpuweb/gpuweb#1160 (61 comments by @ben-clayton, @dneto0, @greggman, @jimblandy, @kainino0x, @kdashg, @litherum, @markusicu, @mehmetoguzderin)
-
Binding model in WGSL does not respect encapsulation gpuweb/gpuweb#2426 (16 comments by @ben-clayton, @dneto0, @kainino0x, @kdashg, @kvark, @litherum, @magcius)
-
Atomic<T> vs Untyped Atomics gpuweb/gpuweb#2377 (16 comments by @amarpMSFT, @dneto0, @kvark, @litherum, @m-schuetz, @raphlinus)
-
Sparse colorAttachments in a render pass gpuweb/gpuweb#1250 (10 comments by @kainino0x, @Kangz, @kvark)
-
Rules about "break" inside "continuing" are strange gpuweb/gpuweb#1867 (7 comments by @dneto0, @jimblandy, @Kangz, @kdashg, @kvark)
-
Make GPUDevice.destroy() a queued operation gpuweb/gpuweb#2599 (7 comments by @greggman, @kainino0x, @Kangz)
-
Proposal: Make `;` optional gpuweb/gpuweb#2577 (7 comments by @ben-clayton, @jimblandy, @jrprice, @magcius)
-
Proposal of FP16 WGSL extension gpuweb/gpuweb#2512 (6 comments by @dneto0, @jiangzhaoming, @kdashg, @kvark, @litherum)
-
need schema for features still under development vs. finalized gpuweb/gpuweb#2608 (6 comments by @alan-baker, @dneto0, @kainino0x, @mehmetoguzderin)
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094 (5 comments by @kainino0x, @kdashg, @litherum)
-
Consider removing GPUPipelineLayout gpuweb/gpuweb#2550 (5 comments by @kainino0x, @Kangz, @kdashg, @kvark, @litherum)
-
GPURenderBundleEncoderDescriptor needs maxCommandCount if it is to be implementable on MTLIndirectCommandBuffer gpuweb/gpuweb#2612 (5 comments by @austinEng, @kainino0x, @litherum)
-
Should `clamp` be only defined for `low` <= `high`? gpuweb/gpuweb#2557 (5 comments by @alan-baker, @dneto0, @kdashg, @kvark)
-
Dynamic indexing in Uniform address space gpuweb/gpuweb#2559 (5 comments by @ben-clayton, @dneto0, @litherum)
-
Proposal: Make `(` `)` optional for `if`, `switch` and `for` gpuweb/gpuweb#2575 (5 comments by @dneto0, @kdashg, @munrocket, @RobinMorisset)
-
Binary WGSL format in createShaderModule() WebAPI gpuweb/gpuweb#2571 (5 comments by @dneto0, @kvark, @munrocket)
-
Add a MatrixInverse builtin? gpuweb/gpuweb#2580 (5 comments by @dneto0, @kainino0x, @kdashg, @magcius)
-
Short i18n review for WebGPU / WGSL gpuweb/gpuweb#2546 (5 comments by @kainino0x, @Kangz, @mehmetoguzderin, @toji)
-
Rules for shadowing built-in functions are unclear gpuweb/gpuweb#2619 (4 comments by @alan-baker, @dneto0, @jimblandy)
-
Should unreachable-code be an error or warning? gpuweb/gpuweb#2378 (4 comments by @alan-baker, @dneto0, @kdashg)
-
Supporting MSAA without rendering capability? gpuweb/gpuweb#2465 (4 comments by @Jiawei-Shao, @kdashg, @kvark)
-
Consider reserving keywords for a module system gpuweb/gpuweb#2607 (4 comments by @dneto0, @jimblandy)
-
Consider making ShaderModule compilation errors special gpuweb/gpuweb#2119 (4 comments by @kainino0x, @kvark, @toji)
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520 (4 comments by @greggman, @kainino0x)
-
Should GPUDevice.destroy detach all buffer mapping ArrayBuffers? gpuweb/gpuweb#2545 (4 comments by @greggman, @kainino0x, @Kangz)
-
mapSync on Workers - and possibly on the main thread gpuweb/gpuweb#2217 (3 comments by @austinEng, @kainino0x, @mrshannon)
-
rgba8unorm canvas textures and IOSurface gpuweb/gpuweb#2535 (3 comments by @ccameron-chromium, @kainino0x, @litherum)
-
Reconsider adding `while` to the language gpuweb/gpuweb#2578 (3 comments by @dneto0, @kdashg, @RobinMorisset)
-
GPUTextureFormatDescriptor.viewFormats performance hint is nonhelpful gpuweb/gpuweb#2573 (3 comments by @kainino0x, @Kangz, @kvark)
-
Feature request: ignore shader writes to color attachments gpuweb/gpuweb#2060 (3 comments by @Kangz)
-
Streaming implementations and indirect draws/dispatches gpuweb/gpuweb#2189 (2 comments by @austinEng, @kdashg)
-
Consider not requiring braces in some situations gpuweb/gpuweb#2211 (2 comments by @kainino0x, @kdashg)
-
Should struct member declarations use commas instead of semicolons? gpuweb/gpuweb#2587 (2 comments by @greggman, @mehmetoguzderin)
-
Are double brackets necessary for attributes? gpuweb/gpuweb#2243 (2 comments by @mehmetoguzderin, @munrocket)
-
Why are bind group layout fields spelled differently from their WGSL counterparts? gpuweb/gpuweb#2469 (2 comments by @Kangz)
-
preserving spir-v developer trust via careful conversion between SPIR-V and WGSL gpuweb/gpuweb#692 (2 comments by @dneto0, @Kangz)
-
Query texture format caps from the device gpuweb/gpuweb#2565 (2 comments by @brendan-duncan, @Kangz)
-
Feature Request : Quaternions as a local type gpuweb/gpuweb#2542 (2 comments by @dneto0, @sketchpunk)
-
wgsl: Missing specification on modf behavior for negatives. gpuweb/gpuweb#2468 (1 comment by @stegu)
-
Out-of-bounds clamping brings big perf regression on TFJS ML demos. gpuweb/gpuweb#1202 (1 comment by @kdashg)
-
Investigation: kill vs demoteToHelper behavior in native APIs gpuweb/gpuweb#921 (1 comment by @kdashg)
-
context-aware tokenization: Sometimes >> is better parsed as two copies of > (similar for ]] ) gpuweb/gpuweb#2092 (1 comment by @kdashg)
-
Proposal: turn "const" at global level into a real storage class gpuweb/gpuweb#1394 (1 comment by @kdashg)
-
Uniformity analysis: find a way to load from a storage buffer uniformly. gpuweb/gpuweb#2321 (1 comment by @Kangz)
-
WGSL type rules need to say something about aliases, and about attributes gpuweb/gpuweb#1387 (1 comment by @kdashg)
-
WGSL should require storage buffers to be a host-shareable storable type. gpuweb/gpuweb#877 (1 comment by @dneto0)
-
further restrict where a control-flow interruption can be placed gpuweb/gpuweb#2093 (1 comment by @dneto0)
-
wgsl: limiting undefined behaviour; subdividing "dynamic error" gpuweb/gpuweb#1844 (1 comment by @dneto0)
-
[wgsl] Define the interface of an entry point in a WGSL program gpuweb/gpuweb#774 (1 comment by @dneto0)
-
SPIR-V doesn't grant reconvergence for multiple returns gpuweb/gpuweb#1770 (1 comment by @RobinMorisset)
-
Metal primitive topology required for the shader compilation hint gpuweb/gpuweb#2574 (1 comment by @litherum)
-
can pipeline creation fail in the driver compiler? particularly due to shader compilation failure gpuweb/gpuweb#1872 (1 comment by @Kangz)
-
MSL question: can mem_flags be combined in threadgroup_barrier? gpuweb/gpuweb#1547 (1 comment by @Kangz)
-
Document behaviour when uniformity requirement can't be proven gpuweb/gpuweb#340 (1 comment by @Kangz)
-
should shaders be checked for execution-uniformity (e.g. barrier only in uniform control flow) gpuweb/gpuweb#264 (1 comment by @Kangz)
-
shader programming model: permitted memory orders on control barriers gpuweb/gpuweb#232 (1 comment by @Kangz)
-
-
39 pull requests opened.
-
Validate query index overwrite in timestampWrites of render pass gpuweb/gpuweb#2627, by @haoxli.
-
wgsl: Cleanup the internal memory layout of vector using SizeOf gpuweb/gpuweb#2626, by @jiangzhaoming.
-
Correct the definition of fragment-stage position builtin gpuweb/gpuweb#2625, by @surma.
-
wgsl: Cleanup internal layout of matrix type gpuweb/gpuweb#2624, by @jiangzhaoming.
-
[editorial] Fix uniformity table widths gpuweb/gpuweb#2623, by @alan-baker.
-
Allow statically unreachable code gpuweb/gpuweb#2622, by @alan-baker.
-
Clarify when a built-in function name can be redefined gpuweb/gpuweb#2621, by @dneto0.
-
Add break-if as optional at end of continuing gpuweb/gpuweb#2618, by @dneto0.
-
wgsl: Reserve words from common programming languages gpuweb/gpuweb#2617, by @dneto0.
-
WGSL style guide: in progress extensions developed outside main spec gpuweb/gpuweb#2616, by @dneto0.
-
Fix a typo in FP32 internal layout gpuweb/gpuweb#2615, by @jiangzhaoming.
-
Update override examples gpuweb/gpuweb#2614, by @alan-baker.
-
wgsl: Rename `firstBitHigh` and `firstBitLow` gpuweb/gpuweb#2611, by @ben-clayton.
-
wgsl: Fix typos gpuweb/gpuweb#2610, by @ben-clayton.
-
wgsl: Describe experimental vs. complete extensions gpuweb/gpuweb#2609, by @dneto0.
-
wgsl: reserve 'std', 'wgsl' gpuweb/gpuweb#2606, by @dneto0.
-
Add importExternalTexture examples with rAF/rVFC rendering gpuweb/gpuweb#2605, by @kainino0x.
-
`createSampler` creates a `GPUSampler` gpuweb/gpuweb#2604, by @bollenberger.
-
Disallow renderable 3D textures gpuweb/gpuweb#2603, by @kvark.
-
Fix a typo in a note "applicaitions". gpuweb/gpuweb#2602, by @bollenberger.
-
Fix typo in "validating texture copy range" argument gpuweb/gpuweb#2596, by @kvark.
-
Allow Unicode Identifiers gpuweb/gpuweb#2595, by @mehmetoguzderin.
-
Remove SPIR-V mappings gpuweb/gpuweb#2594, by @alan-baker.
-
Experiment with using canvas.getContext() for configuration gpuweb/gpuweb#2593, by @toji.
-
Better separate Security and Privacy sections gpuweb/gpuweb#2592, by @toji.
-
Add while loop gpuweb/gpuweb#2590, by @dneto0.
-
if,switch param does not require parentheses gpuweb/gpuweb#2585, by @dneto0.
-
Add isReady(), remove create*PipelineAsync() gpuweb/gpuweb#2583, by @toji.
-
Add acosh, asinh, atanh builtin functions gpuweb/gpuweb#2581, by @dneto0.
-
wgsl: reserve `demote` and `demote_to_helper` gpuweb/gpuweb#2579, by @ben-clayton.
-
Render -- as is gpuweb/gpuweb#2576, by @mehmetoguzderin.
-
Remove createPipelineLayout, take layout "requirements" by value gpuweb/gpuweb#2572, by @kainino0x.
-
Fix step() logic gpuweb/gpuweb#2566, by @kvark.
-
WGSL: Replace [SHORTNAME] with WGSL gpuweb/gpuweb#2564, by @Kangz.
-
Add optional trailing comma for the attribute syntax. gpuweb/gpuweb#2563, by @Kangz.
-
Allow sparse color attachments and ignored FS outputs gpuweb/gpuweb#2562, by @Kangz.
-
Rename GPUTextureSampleTypes to match WGSL gpuweb/gpuweb#2561, by @Kangz.
-
s/endPass/end/ for pass encoders gpuweb/gpuweb#2560, by @Kangz.
-
Rename 'override' to 'param' gpuweb/gpuweb#2556, by @kvark.
-
-
32 pull requests merged.
-
Add break-if as optional at end of continuing gpuweb/gpuweb#2618.
-
[editorial] Fix uniformity table widths gpuweb/gpuweb#2623.
-
wgsl: Cleanup internal layout of matrix type gpuweb/gpuweb#2624.
-
Clarify when a built-in function name can be redefined gpuweb/gpuweb#2621.
-
WGSL style guide: in progress extensions developed outside main spec gpuweb/gpuweb#2616.
-
Fix a typo in FP32 internal layout gpuweb/gpuweb#2615.
-
Update override examples gpuweb/gpuweb#2614.
-
wgsl: Rename `firstBitHigh` and `firstBitLow` gpuweb/gpuweb#2611.
-
wgsl: Fix typos gpuweb/gpuweb#2610.
-
wgsl: reserve 'std', 'wgsl' gpuweb/gpuweb#2606.
-
Extend lifetime of GPUExternalTexture imported from video element gpuweb/gpuweb#2302.
-
`createSampler` creates a `GPUSampler` gpuweb/gpuweb#2604.
-
Disallow renderable 3D textures gpuweb/gpuweb#2603.
-
Fix a typo in a note "applicaitions". gpuweb/gpuweb#2602.
-
Fix typo in "validating texture copy range" argument gpuweb/gpuweb#2596.
-
Complete the Errors & Debugging section gpuweb/gpuweb#2506.
-
Remove SPIR-V mappings gpuweb/gpuweb#2594.
-
Better separate Security and Privacy sections gpuweb/gpuweb#2592.
-
Render -- as is gpuweb/gpuweb#2576.
-
Allow sparse color attachments and ignored FS outputs gpuweb/gpuweb#2562.
-
Add while loop gpuweb/gpuweb#2590.
-
if,switch param does not require parentheses gpuweb/gpuweb#2585.
-
wgsl: reserve `demote` and `demote_to_helper` gpuweb/gpuweb#2579.
-
Add optional trailing comma for the attribute syntax. gpuweb/gpuweb#2563.
-
Rework "validating texture copy range" for 1D/3D textures. gpuweb/gpuweb#2548.
-
Fix canvas resizing example gpuweb/gpuweb#2549.
-
s/endPass/end/ for pass encoders gpuweb/gpuweb#2560.
-
Relax vertex stride requirements gpuweb/gpuweb#2554.
-
Fix step() logic gpuweb/gpuweb#2566.
-
WGSL: Replace [SHORTNAME] with WGSL gpuweb/gpuweb#2564.
-
Vectors consist of components gpuweb/gpuweb#2552.
-
Add the uniformity analysis to the WGSL spec gpuweb/gpuweb#1571.
-
-
6 pull requests closed.
-
Add isReady(), remove create*PipelineAsync() gpuweb/gpuweb#2583.
-
wgsl: Describe experimental vs. complete extensions gpuweb/gpuweb#2609.
-
Rename GPUTextureSampleTypes to match WGSL gpuweb/gpuweb#2561.
-
Add script to extract grammar from WGSL spec gpuweb/gpuweb#1062.
-
Rename 'override' to 'param' gpuweb/gpuweb#2556.
-
wgsl: replace stride attribute with array type parameter gpuweb/gpuweb#2103.
-
-
40 pull requests received 208 new comments.
-
Allow Unicode Identifiers gpuweb/gpuweb#2595 (26 comments by @greggman, @kainino0x, @kdashg, @litherum, @mehmetoguzderin)
-
Add the uniformity analysis to the WGSL spec gpuweb/gpuweb#1571 (22 comments by @kdashg, @kvark, @RobinMorisset)
-
Remove createPipelineLayout, take layout "requirements" by value gpuweb/gpuweb#2572 (19 comments by @kainino0x, @Kangz, @kvark, @magcius, @toji)
-
Rely on canvas size instead of allowing size to be configured gpuweb/gpuweb#2416 (17 comments by @annevk, @kainino0x, @kdashg, @kvark, @toji)
-
Add isReady(), remove create*PipelineAsync() gpuweb/gpuweb#2583 (12 comments by @austinEng, @kainino0x, @Kangz, @kvark, @toji)
-
Allow sparse color attachments and ignored FS outputs gpuweb/gpuweb#2562 (9 comments by @kainino0x, @Kangz, @kvark, @toji)
-
Literal unification, bottom-up gpuweb/gpuweb#2227 (8 comments by @dneto0, @jimblandy, @kdashg, @kvark)
-
Add break-if as optional at end of continuing gpuweb/gpuweb#2618 (7 comments by @alan-baker, @dneto0)
-
Add optional trailing comma for the attribute syntax. gpuweb/gpuweb#2563 (6 comments by @Kangz, @kdashg, @kvark, @mehmetoguzderin)
-
Add acosh, asinh, atanh builtin functions gpuweb/gpuweb#2581 (6 comments by @alan-baker, @dneto0, @kvark)
-
Experiment with using canvas.getContext() for configuration gpuweb/gpuweb#2593 (4 comments by @annevk, @kvark, @toji)
-
Add script to extract grammar from WGSL spec gpuweb/gpuweb#1062 (4 comments by @dneto0, @mehmetoguzderin)
-
Correct the definition of fragment-stage position builtin gpuweb/gpuweb#2625 (3 comments by @kainino0x, @magcius, @surma)
-
Rename GPUTextureSampleTypes to match WGSL gpuweb/gpuweb#2561 (3 comments by @kainino0x, @Kangz, @kvark)
-
Rework "validating texture copy range" for 1D/3D textures. gpuweb/gpuweb#2548 (3 comments by @Kangz, @kvark)
-
Clarify when a built-in function name can be redefined gpuweb/gpuweb#2621 (2 comments by @dneto0, @jimblandy)
-
wgsl: Describe experimental vs. complete extensions gpuweb/gpuweb#2609 (2 comments by @dneto0, @litherum)
-
wgsl: Rename `firstBitHigh` and `firstBitLow` gpuweb/gpuweb#2611 (2 comments by @ben-clayton, @dneto0)
-
Extend lifetime of GPUExternalTexture imported from video element gpuweb/gpuweb#2302 (1 comment by @kdashg)
-
WGSL style guide: in progress extensions developed outside main spec gpuweb/gpuweb#2616 (1 comment by @dneto0)
-
gpuweb/cts
-
15 issues opened.
-
memory model tests: shuffle_groups should be read-only storage buffer gpuweb/cts#1015, by @dneto0.
-
val: encoding,render_bundle gpuweb/cts#1011, by @kainino0x.
-
val: encoding,beginRenderPass gpuweb/cts#1010, by @kainino0x.
-
val: cmds,buffer_texture_copies should move to image_copy gpuweb/cts#1009, by @kainino0x.
-
op: rendering,robust_access_index gpuweb/cts#1008, by @kainino0x.
-
val: cmds,render,draw unimplemented tests other than vertex access gpuweb/cts#1007, by @kainino0x.
-
op: labels gpuweb/cts#1006, by @kainino0x.
-
reftests: test more canvas usages gpuweb/cts#1000, by @austinEng.
-
should tests use isolated random number generators? gpuweb/cts#992, by @dneto0.
-
op: Allow bit patterns for the same value in copyTextureToTexture gpuweb/cts#973, by @Kangz.
-
op: device,lost:same_object gpuweb/cts#971, by @lokokung.
-
op: device,lost:not_lost_on_gc gpuweb/cts#970, by @lokokung.
-
op: lost_on_destroy gpuweb/cts#969, by @lokokung.
-
Add comprehensive memory model tests gpuweb/cts#963, by @reeselevine.
-
-
17 issues closed.
-
stress: device allocation gpuweb/cts#949.
-
reftests: test more canvas usages gpuweb/cts#1000.
-
val: queue,buffer_mapped gpuweb/cts#903.
-
val: queue,destroyed,texture gpuweb/cts#904.
-
Expand tests for 1D textures. gpuweb/cts#872.
-
op: texture_zero 1d textures gpuweb/cts#875.
-
op: copyTextureToTexture 3d textures gpuweb/cts#894.
-
op: copyTextureToTexture 1d textures gpuweb/cts#874.
-
op: lost_on_destroy gpuweb/cts#969.
-
op: device,lost:same_object gpuweb/cts#971.
-
op: device,lost:not_lost_on_gc gpuweb/cts#970.
-
val: image_copy 3d textures gpuweb/cts#893.
-
val: copyTextureToTexture 3d textures gpuweb/cts#892.
-
val: copyTextureToTexture 1d textures gpuweb/cts#876.
-
1 issue reopened.
-
11 issues received 17 new comments.
-
val: beginRenderPass resolveTarget format must support resolve gpuweb/cts#940 (3 comments by @Jiawei-Shao, @kainino0x)
-
should tests use isolated random number generators? gpuweb/cts#992 (3 comments by @dneto0, @kainino0x)
-
op: render_pipeline and compute_pipeline entry_point_name tests gpuweb/cts#895 (2 comments by @kainino0x)
-
memory model tests: shuffle_groups should be read-only storage buffer gpuweb/cts#1015 (1 comment by @dneto0)
-
val: unimplemented createRenderPipeline and createRenderPipelineAsync validation tests gpuweb/cts#982 (1 comment by @kainino0x)
-
-
47 pull requests opened.
-
Implement `atan2` test gpuweb/cts#1023, by @zoddicus.
-
Add additional memory model tests and refactor setup code gpuweb/cts#1022, by @reeselevine.
-
Add validation tests on resolve target formats gpuweb/cts#1021, by @Jiawei-Shao.
-
Fix tests about multisampled textures based on latest SPEC changes gpuweb/cts#1020, by @Jiawei-Shao.
-
webgpu: Remove use of `@stride` for builtin tests gpuweb/cts#1019, by @ben-clayton.
-
replace expired origin trial tokens gpuweb/cts#1018, by @kainino0x.
-
Further split blend_min_max test cases to avoid timing out gpuweb/cts#1017, by @shrekshao.
-
memory model shuffle_workgroups shader var should be read-only gpuweb/cts#1016, by @dneto0.
-
Await on runBufferDependencyTest, copyBufferToBuffer gpuweb/cts#1014, by @shrekshao.
-
Await on runBufferDependencyTest gpuweb/cts#1013, by @shrekshao.
-
wgsl: Add `insertBits` tests gpuweb/cts#1012, by @ben-clayton.
-
wgsl: Add `extractBits` tests gpuweb/cts#1005, by @ben-clayton.
-
wgsl: Add `firstLeadingBit` tests gpuweb/cts#1004, by @ben-clayton.
-
wgsl: Add `firstTrailingBit` tests gpuweb/cts#1003, by @ben-clayton.
-
wgsl: Add `countTrailingZeros` tests gpuweb/cts#1002, by @ben-clayton.
-
wgsl: Add `countLeadingZeros` tests gpuweb/cts#1001, by @ben-clayton.
-
Add rgba16float support for canvas context gpuweb/cts#999, by @jchen10.
-
Throw descriptive error if query doesn't match any cases gpuweb/cts#998, by @austinEng.
-
Review and expand validation,queue,buffer_mapped gpuweb/cts#997, by @toji.
-
Destroyed texture submitting validation test review and implementation gpuweb/cts#996, by @jiangzhaoming.
-
Remove CWD check from cmdline.ts gpuweb/cts#994, by @kainino0x.
-
Add initial memory model tests gpuweb/cts#991, by @reeselevine.
-
Implement validation tests on pipeline_output_targets,format_blendable gpuweb/cts#990, by @Jiawei-Shao.
-
Make cmdline.ts runnable when compiled as cmdline.js as well gpuweb/cts#989, by @austinEng.
-
Swapchain format gpuweb/cts#988, by @jchen10.
-
Partially implement memory_sync,texture,* gpuweb/cts#987, by @austinEng.
-
Swapchain format gpuweb/cts#986, by @jchen10.
-
Move memory_sync,texture,* tests to one file gpuweb/cts#985, by @austinEng.
-
Update to latest WebGPU CTS API gpuweb/cts#984, by @ben-clayton.
-
Implement validation tests on pipeline_output_targets,blend_min_max gpuweb/cts#983, by @Jiawei-Shao.
-
Rewrite the standalone test page with React gpuweb/cts#981, by @austinEng.
-
Fix createView tests for 1D textures gpuweb/cts#980, by @Kangz.
-
texture_zero_init: Expand for 1D textures. gpuweb/cts#979, by @Kangz.
-
Extend copyB2T/copyT2B/WriteTexture tests for 1D/3D textures gpuweb/cts#978, by @Kangz.
-
Adds destroy validation tests for some APIs. gpuweb/cts#977, by @lokokung.
-
Expand copyT2T operation tests to cover 1D and 3D textures. gpuweb/cts#976, by @Kangz.
-
Add `clamp` tests gpuweb/cts#975, by @zoddicus.
-
Remove the last use of readonly-storage-textures. gpuweb/cts#974, by @Kangz.
-
Replace WGSL ignore() with phony assignment. gpuweb/cts#972, by @Kangz.
-
Revert "Re-order some more test parameters for test splitting." gpuweb/cts#968, by @lokokung.
-
Re-order some more test parameters for test splitting. gpuweb/cts#967, by @lokokung.
-
Implement `max` tests gpuweb/cts#966, by @zoddicus.
-
Refactor integer min tests to use generator for tests cases gpuweb/cts#965, by @zoddicus.
-
Add initial test plans for memory model gpuweb/cts#964, by @reeselevine.
-
Re-order test parameters so that splitting it is cleaner. gpuweb/cts#962, by @lokokung.
-
-
45 pull requests merged.
-
Fix tests about multisampled textures based on latest SPEC changes gpuweb/cts#1020.
-
webgpu: Remove use of `@stride` for builtin tests gpuweb/cts#1019.
-
wgsl: Add `insertBits` tests gpuweb/cts#1012.
-
wgsl: Add `extractBits` tests gpuweb/cts#1005.
-
wgsl: Add `firstLeadingBit` tests gpuweb/cts#1004.
-
wgsl: Add `firstTrailingBit` tests gpuweb/cts#1003.
-
wgsl: Add `countTrailingZeros` tests gpuweb/cts#1002.
-
wgsl: Add `countLeadingZeros` tests gpuweb/cts#1001.
-
memory model shuffle_workgroups shader var should be read-only gpuweb/cts#1016.
-
replace expired origin trial tokens gpuweb/cts#1018.
-
Await on runBufferDependencyTest, copyBufferToBuffer gpuweb/cts#1014.
-
Partially implement memory_sync,texture,* gpuweb/cts#987.
-
Await on runBufferDependencyTest gpuweb/cts#1013.
-
Add rgba16float support for canvas context gpuweb/cts#999.
-
Review and expand validation,queue,buffer_mapped gpuweb/cts#997.
-
Add initial memory model tests gpuweb/cts#991.
-
Throw descriptive error if query doesn't match any cases gpuweb/cts#998.
-
Remove CWD check from cmdline.ts gpuweb/cts#994.
-
Add rgba8unorm support for swapchain format gpuweb/cts#952.
-
Swapchain format gpuweb/cts#988.
-
Swapchain format gpuweb/cts#986.
-
Move memory_sync,texture,* tests to one file gpuweb/cts#985.
-
Plan api,operation,memory_sync,texture,* gpuweb/cts#954.
-
Add initial test plans for memory model gpuweb/cts#964.
-
Adds stress tests for device destroy. gpuweb/cts#961.
-
Add `clamp` tests gpuweb/cts#975.
-
Extend copyB2T/copyT2B/WriteTexture tests for 1D/3D textures gpuweb/cts#978.
-
Fix createView tests for 1D textures gpuweb/cts#980.
-
texture_zero_init: Expand for 1D textures. gpuweb/cts#979.
-
Expand copyT2T operation tests to cover 1D and 3D textures. gpuweb/cts#976.
-
Remove the last use of readonly-storage-textures. gpuweb/cts#974.
-
Replace WGSL ignore() with phony assignment. gpuweb/cts#972.
-
Refactor integer min tests to use generator for tests cases gpuweb/cts#965.
-
Implement `max` tests gpuweb/cts#966.
-
Revert "Re-order some more test parameters for test splitting." gpuweb/cts#968.
-
Re-order some more test parameters for test splitting. gpuweb/cts#967.
-
Update image copy validation tests for 1D / 3D textures gpuweb/cts#960.
-
Re-order test parameters so that splitting it is cleaner. gpuweb/cts#962.
-
-
2 pull requests closed.
-
Further split blend_min_max test cases to avoid timing out gpuweb/cts#1017.
-
Make cmdline.ts runnable when compiled as cmdline.js as well gpuweb/cts#989.
-
-
26 pull requests received 252 new comments.
-
Adds destroy validation tests for some APIs. gpuweb/cts#977 (40 comments by @austinEng, @kainino0x, @lokokung)
-
Add initial test plans for memory model gpuweb/cts#964 (14 comments by @dneto0, @reeselevine, @zoddicus)
-
Destroyed texture submitting validation test review and implementation gpuweb/cts#996 (11 comments by @austinEng, @jiangzhaoming, @kainino0x)
-
Expand copyT2T operation tests to cover 1D and 3D textures. gpuweb/cts#976 (11 comments by @austinEng, @Kangz)
-
Extend copyB2T/copyT2B/WriteTexture tests for 1D/3D textures gpuweb/cts#978 (8 comments by @austinEng, @Kangz)
-
Make cmdline.ts runnable when compiled as cmdline.js as well gpuweb/cts#989 (6 comments by @austinEng, @kainino0x)
-
Reland "Implement validation tests on pipeline_output_targets,format_blendable (#990)" gpuweb/cts#995 (5 comments by @austinEng, @Jiawei-Shao)
-
Implement validation tests on pipeline_output_targets,blend_min_max gpuweb/cts#983 (3 comments by @Kangz)
-
Refactor integer min tests to use generator for tests cases gpuweb/cts#965 (3 comments by @ben-clayton, @dneto0, @zoddicus)
-
memory model shuffle_workgroups shader var should be read-only gpuweb/cts#1016 (2 comments by @ben-clayton, @dneto0)
-
Revert "Implement validation tests on pipeline_output_targets,format_blendable" gpuweb/cts#993 (1 comment by @kainino0x)
-
wgsl: Test that integral user-defined IO requires flat interpolation gpuweb/cts#951 (1 comment by @zoddicus)
TR publications
-
32 documents published.
-
WebGPU published 16 times as a Working Draft. History of WebGPU
-
WebGPU Shading Language published 16 times as a Working Draft. History of WebGPU Shading Language
-
Participants
-
4 participants joined.
-
Todd Kosloff (OPPO) joined.
-
he feng (OPPO) joined.
-
Xunran Ding (OPPO) joined.
-
Weixin Dong (OPPO) joined.
-
-
1 organization joined.
-
OPPO joined.
-
GitHub
gpuweb/gpuweb
-
38 issues opened.
-
Fragment Built-in Output Value for Stencil [Feature Request] gpuweb/gpuweb#2553, by @eurigilberto.
-
Consider removing GPUPipelineLayout gpuweb/gpuweb#2550, by @kainino0x.
-
Separate the privacy and security consideration sections of the spec. gpuweb/gpuweb#2547, by @Kangz.
-
Short i18n review for WebGPU / WGSL gpuweb/gpuweb#2546, by @Kangz.
-
Should GPUDevice.destroy detach all buffer mapping ArrayBuffers? gpuweb/gpuweb#2545, by @Kangz.
-
Feature Request : Quaternions as a local type gpuweb/gpuweb#2542, by @sketchpunk.
-
rgba8unorm canvas textures and IOSurface gpuweb/gpuweb#2535, by @kainino0x.
-
Invalid example for requiredFeatures gpuweb/gpuweb#2532, by @pyoor.
-
Seek horizontal reviews gpuweb/gpuweb#2531, by @Kangz.
-
Behavior of discarding everything gpuweb/gpuweb#2523, by @kvark.
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520, by @greggman.
-
Spec guidance on when/where error messages appear gpuweb/gpuweb#2516, by @kainino0x.
-
Proposal of FP16 WGSL extension gpuweb/gpuweb#2512, by @jiangzhaoming.
-
Parameter order in writeBuffer and copyBufferToBuffer gpuweb/gpuweb#2511, by @terier.
-
Clarify the rules for scoping/shadowing gpuweb/gpuweb#2510, by @RobinMorisset.
-
misleading chrome error for bind groups gpuweb/gpuweb#2508, by @martin19.
-
multi pass box blur compute shader gpuweb/gpuweb#2507, by @martin19.
-
constrain attribute alignment value to divide device alignment limits gpuweb/gpuweb#2505, by @dneto0.
-
editorial: Add table of correspondences for builtin functions in common shader languages gpuweb/gpuweb#2502, by @dneto0.
-
Determine whether getBindGroupLayout returns a new object or not gpuweb/gpuweb#2501, by @kainino0x.
-
Investigation: how WebNN / WebGPU interop could be happening gpuweb/gpuweb#2500, by @huningxin.
-
Where do we spec WebGPU-WebCodecs interaction? gpuweb/gpuweb#2498, by @kainino0x.
-
Structure/array size rules are repeated in the spec gpuweb/gpuweb#2497, by @kvark.
-
remove stride attribute gpuweb/gpuweb#2493, by @dneto0.
-
struct decl should not have to be followed by a semicolon gpuweb/gpuweb#2492, by @dneto0.
-
Require that 1D texture have a single mipmap. gpuweb/gpuweb#2490, by @Kangz.
-
depthOrArrayLayers gpuweb/gpuweb#2489, by @unconed.
-
in certain contexts, an identifier does not resolve to its declared object gpuweb/gpuweb#2486, by @dneto0.
-
Dynamic texture selection is impossible for sample() calls gpuweb/gpuweb#2482, by @litherum.
-
column_major and row_major matrix annotations, WRT pre/post matrix multiplication gpuweb/gpuweb#2481, by @JasonHuang3D.
-
Handling Errors Synchronously gpuweb/gpuweb#2478, by @greggman.
-
editorial: remove "A name cannot be used before it is declared." gpuweb/gpuweb#2477, by @dneto0.
-
wgsl: only define texel format names valid for storage textures in WGSL gpuweb/gpuweb#2473, by @dneto0.
-
Why is WGSL verbose? gpuweb/gpuweb#2471, by @Jerrody.
-
Consider removing or reworking implicit pipeline layouts gpuweb/gpuweb#2470, by @magcius.
-
Why are bind group layout fields spelled differently from their WGSL counterparts? gpuweb/gpuweb#2469, by @magcius.
-
wgsl: Missing specification on modf behavior for negatives. gpuweb/gpuweb#2468, by @Axiverse.
-
-
37 issues closed.
-
allow array size that is a pipeline-overrideable constant gpuweb/gpuweb#1431.
-
Removing the COPY* texture usages gpuweb/gpuweb#2196.
-
depthOrArrayLayers gpuweb/gpuweb#2489.
-
Invalid example for requiredFeatures gpuweb/gpuweb#2532.
-
"Storage class" term is unnecessarily confusing gpuweb/gpuweb#2440.
-
Refactor API for load/store OP and read-only attachments gpuweb/gpuweb#2131.
-
Reconsider GPULoadOp "clear"/"zero" gpuweb/gpuweb#1377.
-
Should GPUCompliationIssue.messages be an ObservableArray? gpuweb/gpuweb#2447.
-
Behavior of endPass() and pass commands after pass has ended gpuweb/gpuweb#2464.
-
Parameter order in writeBuffer and copyBufferToBuffer gpuweb/gpuweb#2511.
-
Structure/array size rules are repeated in the spec gpuweb/gpuweb#2497.
-
How to measure the shader compilation time in js gpuweb/gpuweb#2236.
-
Storage variables shouldn't be required to be structs gpuweb/gpuweb#2188.
-
Are double brackets necessary for attributes? gpuweb/gpuweb#2243.
-
struct decl should not have to be followed by a semicolon gpuweb/gpuweb#2492.
-
when does applying an attribute to a type result in a different type gpuweb/gpuweb#1534.
-
remove stride attribute gpuweb/gpuweb#2493.
-
misleading chrome error for bind groups gpuweb/gpuweb#2508.
-
Clarify the rules for scoping/shadowing gpuweb/gpuweb#2510.
-
multi pass box blur compute shader gpuweb/gpuweb#2507.
-
Consider making some optional members also nullable gpuweb/gpuweb#1400.
-
Determine GPUExternalTexture lifetime gpuweb/gpuweb#2124.
-
Considerations for timing attacks using timestamp queries gpuweb/gpuweb#2218.
-
Select operator arguments are in a non-intuitive order gpuweb/gpuweb#2242.
-
Allow float modulo for scalar-vector and vector-scalar gpuweb/gpuweb#2450.
-
wgsl: support bitfield insertion and extraction gpuweb/gpuweb#2129.
-
wgsl: finding most-significant bit and least-significant bit gpuweb/gpuweb#2130.
-
Do we need a separate RESOLVE texture usage bit? gpuweb/gpuweb#2408.
-
Require that 1D texture have a single mipmap. gpuweb/gpuweb#2490.
-
spec: forbid infinite loops gpuweb/gpuweb#2414.
-
editorial: remove "A name cannot be used before it is declared." gpuweb/gpuweb#2477.
-
wgsl editorial: Builtin variables are not "variables" gpuweb/gpuweb#2445.
-
wgsl: only define texel format names valid for storage textures in WGSL gpuweb/gpuweb#2473.
-
Why is WGSL verbose? gpuweb/gpuweb#2471.
-
image format names, 'read', 'read_write' should not be keywords gpuweb/gpuweb#2428.
-
-
79 issues received 253 new comments.
-
column_major and row_major matrix annotations, WRT pre/post matrix multiplication gpuweb/gpuweb#2481 (15 comments by @ben-clayton, @greggman, @JasonHuang3D, @Kangz, @kdashg, @kvark, @magcius, @X-OldGentleMan)
-
Should unreachable-code be an error or warning? gpuweb/gpuweb#2378 (14 comments by @alan-baker, @dneto0, @kdashg, @litherum, @RobinMorisset)
-
Handling Errors Synchronously gpuweb/gpuweb#2478 (13 comments by @greggman, @kainino0x, @Kangz, @kvark, @magcius)
-
Dynamic texture selection is impossible for sample() calls gpuweb/gpuweb#2482 (11 comments by @alan-baker, @dneto0, @kvark, @litherum, @magcius)
-
Consider removing or reworking implicit pipeline layouts gpuweb/gpuweb#2470 (8 comments by @brendan-duncan, @kainino0x, @Kangz, @m-schuetz, @magcius)
-
Are double brackets necessary for attributes? gpuweb/gpuweb#2243 (8 comments by @aloucks, @ben-clayton, @dneto0, @kdashg, @litherum, @mehmetoguzderin, @RobinMorisset)
-
Should GPUDevice.destroy detach all buffer mapping ArrayBuffers? gpuweb/gpuweb#2545 (7 comments by @kainino0x, @Kangz, @lokokung)
-
constrain attribute alignment value to divide device alignment limits gpuweb/gpuweb#2505 (7 comments by @dneto0, @kainino0x, @Kangz)
-
Should objects have pointers back to their parents? gpuweb/gpuweb#1497 (7 comments by @kainino0x, @Kangz, @kdashg, @kvark, @toji)
-
Should objects expose the descriptor they were created with? gpuweb/gpuweb#1498 (7 comments by @AshleyScirra, @kainino0x, @Kangz, @kdashg, @kvark, @shannon)
-
mapSync on Workers - and possibly on the main thread gpuweb/gpuweb#2217 (6 comments by @kainino0x, @mrshannon)
-
Why is WGSL verbose? gpuweb/gpuweb#2471 (6 comments by @ben-clayton, @Jerrody, @kvark, @litherum, @mehmetoguzderin)
-
copyExternalImageToTexture/importExternalTexture with destination in linear color space gpuweb/gpuweb#1715 (6 comments by @donmccurdy, @kainino0x, @kvark)
-
Proposal: Add a usage bit to allow creating a view with different format when creating a texture gpuweb/gpuweb#168 (5 comments by @Kangz, @kdashg, @litherum, @shaoboyan)
-
Investigation: how WebNN / WebGPU interop could be happening gpuweb/gpuweb#2500 (5 comments by @chcunningham, @huningxin, @Kangz)
-
wgsl: support extended integer arithmetic for add, subtract, multiply gpuweb/gpuweb#1565 (4 comments by @Kangz, @kdashg, @litherum)
-
remove stride attribute gpuweb/gpuweb#2493 (4 comments by @kdashg, @kvark, @litherum, @RobinMorisset)
-
Consider making some optional members also nullable gpuweb/gpuweb#1400 (4 comments by @kainino0x, @Kangz)
-
Atomic<T> vs Untyped Atomics gpuweb/gpuweb#2377 (4 comments by @ben-clayton, @kvark, @litherum, @raphlinus)
-
Handling out-of-memory when resizing canvas backing store gpuweb/gpuweb#2520 (3 comments by @greggman, @kainino0x)
-
Behavior of endPass() and pass commands after pass has ended gpuweb/gpuweb#2464 (3 comments by @kainino0x, @Kangz)
-
Parameter order in writeBuffer and copyBufferToBuffer gpuweb/gpuweb#2511 (3 comments by @kainino0x, @terier)
-
Create a proposal for multi-worker gpuweb/gpuweb#354 (3 comments by @austinEng, @greggman, @kainino0x)
-
Where do we spec WebGPU-WebCodecs interaction? gpuweb/gpuweb#2498 (3 comments by @dalecurtis, @kainino0x, @kdashg)
-
Select operator arguments are in a non-intuitive order gpuweb/gpuweb#2242 (3 comments by @ben-clayton, @dneto0, @kdashg)
-
Why are bind group layout fields spelled differently from their WGSL counterparts? gpuweb/gpuweb#2469 (3 comments by @Kangz, @kvark, @magcius)
-
Feature Request : Quaternions as a local type gpuweb/gpuweb#2542 (2 comments by @greggman, @munrocket)
-
Spec guidance on when/where error messages appear gpuweb/gpuweb#2516 (2 comments by @greggman, @kainino0x)
-
Exposing GPU hardware information to web developers in WebGPU gpuweb/gpuweb#2195 (2 comments by @toji)
-
SPIR-V doesn't grant reconvergence for multiple returns gpuweb/gpuweb#1770 (2 comments by @dneto0, @RobinMorisset)
-
struct decl should not have to be followed by a semicolon gpuweb/gpuweb#2492 (2 comments by @kdashg, @kvark)
-
Binding model in WGSL does not respect encapsulation gpuweb/gpuweb#2426 (2 comments by @Kangz, @litherum)
-
Clarify the rules for scoping/shadowing gpuweb/gpuweb#2510 (2 comments by @ben-clayton, @RobinMorisset)
-
Determine whether getBindGroupLayout returns a new object or not gpuweb/gpuweb#2501 (2 comments by @kainino0x, @Kangz)
-
Add gpuCommandEncoder.fillBuffer8() and gpuCommandEncoder.fillBuffer32() gpuweb/gpuweb#2372 (2 comments by @Kangz, @litherum)
-
Allow float modulo for scalar-vector and vector-scalar gpuweb/gpuweb#2450 (2 comments by @dneto0, @kdashg)
-
Fragment Built-in Output Value for Stencil [Feature Request] gpuweb/gpuweb#2553 (1 comment by @dneto0)
-
Fake transfers for getMappedRange-created ArrayBuffers? gpuweb/gpuweb#2072 (1 comment by @kainino0x)
-
Section 3.30 is titled "Expression statement", but actually defines assignment statements gpuweb/gpuweb#758 (1 comment by @kdashg)
-
Should GPUCompliationIssue.messages be an ObservableArray? gpuweb/gpuweb#2447 (1 comment by @kainino0x)
-
Should the API enforce that writable storage buffer bindings don't alias each other? gpuweb/gpuweb#1842 (1 comment by @kainino0x)
-
editorial: Add table of correspondences for builtin functions in common shader languages gpuweb/gpuweb#2502 (1 comment by @mehmetoguzderin)
-
Consider making ShaderModule compilation errors special gpuweb/gpuweb#2119 (1 comment by @kainino0x)
-
Allow copying from any canvas that toDataURL can be called on gpuweb/gpuweb#2373 (1 comment by @Kangz)
-
texture_2d_depth for loading depth values might be unnecessary? gpuweb/gpuweb#2094 (1 comment by @Kangz)
-
Method of ensuring GPUShaderModules can contain MTLLibraries gpuweb/gpuweb#1064 (1 comment by @litherum)
-
-
43 pull requests opened.
-
Relax vertex stride requirements gpuweb/gpuweb#2554, by @kvark.
-
Vectors consist of components gpuweb/gpuweb#2552, by @mehmetoguzderin.
-
Resurrect canvas srgb formats gpuweb/gpuweb#2551, by @kainino0x.
-
Fix canvas resizing example gpuweb/gpuweb#2549, by @bbbbx.
-
Rework "validating texture copy range" for 1D/3D textures. gpuweb/gpuweb#2548, by @Kangz.
-
[editorial] Use consistent capitalisation in section titles gpuweb/gpuweb#2544, by @alan-baker.
-
Remove unnecessary todos gpuweb/gpuweb#2543, by @alan-baker.
-
Add GPUTextureDescriptor viewFormats list gpuweb/gpuweb#2540, by @kainino0x.
-
editorial: discuss parsing ambiguity: type construction, function call gpuweb/gpuweb#2539, by @dneto0.
-
Enforce presence of an initializer for module-scope let gpuweb/gpuweb#2538, by @alan-baker.
-
Move all canvas configure validation to getCurrentTexture gpuweb/gpuweb#2536, by @kainino0x.
-
Explain smoothStep better gpuweb/gpuweb#2534, by @dneto0.
-
Fix clamp arity in error bounds section gpuweb/gpuweb#2533, by @dneto0.
-
wgsl: Fix example's builtin name gpuweb/gpuweb#2530, by @bbbbx.
-
Remove duplicated words gpuweb/gpuweb#2529, by @jiangzhaoming.
-
Rename storage class into address space gpuweb/gpuweb#2524, by @kvark.
-
Refine the supported swapchain formats gpuweb/gpuweb#2522, by @jchen10.
-
editorial: streamline array and structure layout descriptions gpuweb/gpuweb#2521, by @dneto0.
-
Add note explaining how dispatch args and workgroup sizes interact gpuweb/gpuweb#2519, by @toji.
-
integer division corresponds to OpSRem gpuweb/gpuweb#2518, by @dneto0.
-
Switch to `@` for Attributes gpuweb/gpuweb#2517, by @mehmetoguzderin.
-
Update behavior analysis for unconditional discards gpuweb/gpuweb#2515, by @alan-baker.
-
Remove tiny duplicates gpuweb/gpuweb#2514, by @mehmetoguzderin.
-
Fix a typo in packing built-in functions list in WGSL spec gpuweb/gpuweb#2513, by @jiangzhaoming.
-
Complete the Errors & Debugging section gpuweb/gpuweb#2506, by @toji.
-
Fix map/list confusion from #2441 gpuweb/gpuweb#2504, by @kainino0x.
-
Remove stride attribute gpuweb/gpuweb#2503, by @dneto0.
-
Struct decl does not have to end in a semicolon gpuweb/gpuweb#2499, by @dneto0.
-
Change GPUObjectBase.label from nullable to union-with-undefined gpuweb/gpuweb#2496, by @kainino0x.
-
Add mixed vector-scalar float % operator gpuweb/gpuweb#2495, by @dneto0.
-
Mirror descriptor properties on GPU resources gpuweb/gpuweb#2494, by @toji.
-
Require 1D texture mipLevelCount to be 1. gpuweb/gpuweb#2491, by @Kangz.
-
Add simple examples for create, init, and error handling functions gpuweb/gpuweb#2488, by @toji.
-
reworking uniformity analysis PR gpuweb/gpuweb#2487, by @dneto0.
-
Clarify fract gpuweb/gpuweb#2485, by @dneto0.
-
Add pointers back to parent objects for most interfaces gpuweb/gpuweb#2484, by @toji.
-
wgsl: Remove notes about non-ref dynamic indexing gpuweb/gpuweb#2483, by @jrprice.
-
Move Limits section to under "WGSL Program" gpuweb/gpuweb#2480, by @dneto0.
-
Fix declaration-and-scope section for out-of-order decls gpuweb/gpuweb#2479, by @dneto0.
-
[editorial] Rename "built-in variable" -> "built-in value" gpuweb/gpuweb#2476, by @dneto0.
-
Only define image format names usable for storage textures gpuweb/gpuweb#2475, by @dneto0.
-
Remove 'read','read_write','write' as keywords, image formats as keywords gpuweb/gpuweb#2474, by @dneto0.
-
[editorial] wgsl: left shifts are logical gpuweb/gpuweb#2472, by @jrprice.
-
-
42 pull requests merged.
-
Enforce presence of an initializer for module-scope let gpuweb/gpuweb#2538.
-
Add GPUTextureDescriptor viewFormats list gpuweb/gpuweb#2540.
-
Remove unnecessary todos gpuweb/gpuweb#2543.
-
[editorial] Use consistent capitalisation in section titles gpuweb/gpuweb#2544.
-
Fix clamp arity in error bounds section gpuweb/gpuweb#2533.
-
Explain smoothStep better gpuweb/gpuweb#2534.
-
Rename storage class into address space gpuweb/gpuweb#2524.
-
Add override declarations gpuweb/gpuweb#2404.
-
wgsl: detailed semantics of integer division and remainder gpuweb/gpuweb#1830.
-
wgsl: Fix example's builtin name gpuweb/gpuweb#2530.
-
Refine the supported swapchain formats gpuweb/gpuweb#2522.
-
Add note explaining how dispatch args and workgroup sizes interact gpuweb/gpuweb#2519.
-
Separate loadOp and clear values gpuweb/gpuweb#2386.
-
Change GPUObjectBase.label from nullable to union-with-undefined gpuweb/gpuweb#2496.
-
editorial: streamline array and structure layout descriptions gpuweb/gpuweb#2521.
-
Remove duplicated words gpuweb/gpuweb#2529.
-
Store type for buffer does not have to be structure gpuweb/gpuweb#2401.
-
wgsl: float to integer conversion saturates gpuweb/gpuweb#2434.
-
Struct decl does not have to end in a semicolon gpuweb/gpuweb#2499.
-
Switch to `@` for Attributes gpuweb/gpuweb#2517.
-
Remove stride attribute gpuweb/gpuweb#2503.
-
integer division corresponds to OpSRem gpuweb/gpuweb#2518.
-
Remove tiny duplicates gpuweb/gpuweb#2514.
-
Fix a typo in packing built-in functions list in WGSL spec gpuweb/gpuweb#2513.
-
Fix map/list confusion from #2441 gpuweb/gpuweb#2504.
-
Disallow aliasing writable resources gpuweb/gpuweb#2441.
-
wgsl: Remove notes about non-ref dynamic indexing gpuweb/gpuweb#2483.
-
Add mixed vector-scalar float % operator gpuweb/gpuweb#2495.
-
wgsl: Add extractBits, insertBits gpuweb/gpuweb#2466.
-
wsgl: Add bit-finding functions. gpuweb/gpuweb#2467.
-
Add simple examples for create, init, and error handling functions gpuweb/gpuweb#2488.
-
Require 1D texture mipLevelCount to be 1. gpuweb/gpuweb#2491.
-
Officially add Brandon as a spec editor gpuweb/gpuweb#2418.
-
Clarify fract gpuweb/gpuweb#2485.
-
Behaviors: Ban obviously infinite loops gpuweb/gpuweb#2430.
-
Fix declaration-and-scope section for out-of-order decls gpuweb/gpuweb#2479.
-
Move Limits section to under "WGSL Program" gpuweb/gpuweb#2480.
-
[editorial] Rename "built-in variable" -> "built-in value" gpuweb/gpuweb#2476.
-
Only define image format names usable for storage textures gpuweb/gpuweb#2475.
-
Remove 'read','read_write','write' as keywords, image formats as keywords gpuweb/gpuweb#2474.
-
[editorial] wgsl: left shifts are logical gpuweb/gpuweb#2472.
-
-
5 pull requests closed.
-
Add GPUTextureDescriptor option for srgb reinterpretation gpuweb/gpuweb#2336.
-
Issue 1534's array types differ on stride gpuweb/gpuweb#1943.
-
Update behavior analysis for unconditional discards gpuweb/gpuweb#2515.
-
wgsl: OpXMod -> OpXRem for integers gpuweb/gpuweb#2187.
-
Typing of literals gpuweb/gpuweb#2306.
-
-
52 pull requests received 347 new comments.
-
Add the uniformity analysis to the WGSL spec gpuweb/gpuweb#1571 (63 comments by @alan-baker, @dneto0, @kdashg, @RobinMorisset)
-
Literal unification, bottom-up gpuweb/gpuweb#2227 (25 comments by @alan-baker, @ben-clayton, @dneto0, @kdashg)
-
Add simple examples for create, init, and error handling functions gpuweb/gpuweb#2488 (21 comments by @kainino0x, @toji)
-
Rename storage class into address space gpuweb/gpuweb#2524 (18 comments by @dneto0, @kdashg, @kvark, @litherum)
-
Store type for buffer does not have to be structure gpuweb/gpuweb#2401 (17 comments by @alan-baker, @dneto0, @jrprice, @kvark)
-
reworking uniformity analysis PR gpuweb/gpuweb#2487 (16 comments by @dneto0, @Kangz, @RobinMorisset)
-
Mirror descriptor properties on GPU resources gpuweb/gpuweb#2494 (15 comments by @juj, @kainino0x, @Kangz, @toji)
-
editorial: discuss parsing ambiguity: type construction, function call gpuweb/gpuweb#2539 (14 comments by @alan-baker, @dneto0, @mehmetoguzderin)
-
wsgl: Add bit-finding functions. gpuweb/gpuweb#2467 (14 comments by @alan-baker, @ben-clayton, @dneto0, @kdashg)
-
editorial: streamline array and structure layout descriptions gpuweb/gpuweb#2521 (12 comments by @alan-baker, @dneto0)
-
wgsl: detailed semantics of integer division and remainder gpuweb/gpuweb#1830 (7 comments by @dneto0, @kdashg, @kvark)
-
Update behavior analysis for unconditional discards gpuweb/gpuweb#2515 (6 comments by @alan-baker, @dneto0, @RobinMorisset)
-
Rely on canvas size instead of allowing size to be configured gpuweb/gpuweb#2416 (5 comments by @kainino0x, @kvark)
-
Add note explaining how dispatch args and workgroup sizes interact gpuweb/gpuweb#2519 (5 comments by @jrprice, @kainino0x, @toji)
-
wgsl: float to integer conversion saturates gpuweb/gpuweb#2434 (5 comments by @dneto0, @kdashg, @kvark)
-
wgsl: OpXMod -> OpXRem for integers gpuweb/gpuweb#2187 (5 comments by @alan-baker, @ben-clayton, @dneto0, @kdashg)
-
Fix declaration-and-scope section for out-of-order decls gpuweb/gpuweb#2479 (4 comments by @alan-baker, @dneto0)
-
[editorial] Rename "built-in variable" -> "built-in value" gpuweb/gpuweb#2476 (4 comments by @alan-baker, @dneto0)
-
Enforce presence of an initializer for module-scope let gpuweb/gpuweb#2538 (3 comments by @alan-baker, @mehmetoguzderin)
-
Add proposals/ directory and pipeline statistics extension doc gpuweb/gpuweb#2301 (3 comments by @kainino0x, @litherum, @tidoust)
-
Add pointers back to parent objects for most interfaces gpuweb/gpuweb#2484 (3 comments by @kainino0x, @toji)
-
Extend lifetime of GPUExternalTexture imported from video element gpuweb/gpuweb#2302 (2 comments by @kainino0x)
-
Add GPUTextureDescriptor option for srgb reinterpretation gpuweb/gpuweb#2336 (2 comments by @kainino0x)
-
[Process] Add RequirementsForAdditionalFunctionality.md gpuweb/gpuweb#2424 (2 comments by @kainino0x, @litherum)
-
Only require/allow depth/stencil load/store ops when they actually have an effect gpuweb/gpuweb#2387 (2 comments by @kainino0x)
-
Struct decl does not have to end in a semicolon gpuweb/gpuweb#2499 (2 comments by @dneto0, @mehmetoguzderin)
-
Change GPUObjectBase.label from nullable to union-with-undefined gpuweb/gpuweb#2496 (1 comment by @kainino0x)
-
Remove 'read','read_write','write' as keywords, image formats as keywords gpuweb/gpuweb#2474 (1 comment by @dneto0)
-
wgsl: fold expression semantics text into definition of type rule gpuweb/gpuweb#2444 (1 comment by @alan-baker)
-
gpuweb/cts
-
67 issues opened.
-
web_platform: copyExternalImageToTexture test flipY with all sources gpuweb/cts#957, by @shaoboyan.
-
stress: device allocation gpuweb/cts#949, by @lokokung.
-
reftest: rendering via transferControlToOffscreen gpuweb/cts#941, by @kainino0x.
-
val: beginRenderPass resolveTarget format must support resolve gpuweb/cts#940, by @kainino0x.
-
val: required limits must be a power of 2 gpuweb/cts#939, by @kainino0x.
-
val: buffer,create:usage gpuweb/cts#937, by @kainino0x.
-
val: buffer,create:size gpuweb/cts#936, by @kainino0x.
-
op: copyTextureToTexture between srgb and non-srgb gpuweb/cts#935, by @kainino0x.
-
op: test pipeline caching gpuweb/cts#934, by @kainino0x.
-
op: render_pass,storeop2 gpuweb/cts#933, by @kainino0x.
-
How to differentiate between backends due to internal details/rounding gpuweb/cts#930, by @zoddicus.
-
val: mapAsync,state,destroyed gpuweb/cts#929, by @kainino0x.
-
val: attachment_compatibility allow unwritten output if writeMask=0 gpuweb/cts#928, by @kainino0x.
-
val: attachment_compatibility for read-only depth/stencil gpuweb/cts#927, by @kainino0x.
-
op: pipeline,default_layout:layout - default layout is computed correctly gpuweb/cts#924, by @kainino0x.
-
val: resource_usages,texture,in_pass_encoder gpuweb/cts#923, by @kainino0x.
-
web_platform: user-initiated canvas readback results gpuweb/cts#922, by @kainino0x.
-
val: capability_checks,features errors should be exceptions gpuweb/cts#919, by @kainino0x.
-
reftests: test canvas colorSpace gpuweb/cts#918, by @kainino0x.
-
reftests: test canvas compositingAlphaMode gpuweb/cts#917, by @kainino0x.
-
reftests: test more ways to write into canvas texture gpuweb/cts#916, by @kainino0x.
-
reftests: test more canvas texture formats gpuweb/cts#915, by @kainino0x.
-
op: render_pipeline,vertex_only_render_pipeline gpuweb/cts#914, by @kainino0x.
-
val: device_mismatch throughout CTS gpuweb/cts#912, by @kainino0x.
-
op: memory_allocation gpuweb/cts#911, by @kainino0x.
-
web_platform: test copies and imports from WebCodecs VideoFrame gpuweb/cts#909, by @kainino0x.
-
web_platform: context_creation from transferControlToOffscreen gpuweb/cts#908, by @kainino0x.
-
op: texture_view,read/write gpuweb/cts#907, by @kainino0x.
-
val: Resource usage validation for textures/buffers gpuweb/cts#905, by @kainino0x.
-
val: cmds,render,draw vertex access tests gpuweb/cts#906, by @kainino0x.
-
val: queue,destroyed,texture gpuweb/cts#904, by @kainino0x.
-
val: queue,buffer_mapped gpuweb/cts#903, by @kainino0x.
-
op: rendering,draw:default_arguments gpuweb/cts#900, by @kainino0x.
-
Update tests for sparse render attachments if added gpuweb/cts#901, by @kainino0x.
-
op: memory_sync,buffer,ww:multiple_buffers gpuweb/cts#897, by @kainino0x.
-
op: Test that oversized dispatches don't dispatch any threads gpuweb/cts#896, by @kainino0x.
-
op: memory_sync,texture,rw_and_wr gpuweb/cts#899, by @kainino0x.
-
op: memory_sync,texture,ww gpuweb/cts#898, by @kainino0x.
-
op: render_pipeline and compute_pipeline entry_point_name tests gpuweb/cts#895, by @kainino0x.
-
op: copyTextureToTexture 3d textures gpuweb/cts#894, by @kainino0x.
-
val: image_copy 3d textures gpuweb/cts#893, by @kainino0x.
-
val: image_copy gpuweb/cts#891, by @kainino0x.
-
val: copyTextureToTexture 3d textures gpuweb/cts#892, by @kainino0x.
-
op: commmand_buffer,occlusion_queries gpuweb/cts#888, by @kainino0x.
-
val: layout_shader_compat gpuweb/cts#889, by @kainino0x.
-
op: rendering,depth_clip_clamp gpuweb/cts#890, by @kainino0x.
-
op: memory_sync,buffer,rw_and_wr gpuweb/cts#887, by @kainino0x.
-
val: setBindGroup gpuweb/cts#886, by @kainino0x.
-
val: createBindGroupLayout gpuweb/cts#885, by @kainino0x.
-
val: createBindGroup gpuweb/cts#884, by @kainino0x.
-
op: vertex_state,correctness non-number cases and float tolerance gpuweb/cts#883, by @kainino0x.
-
val: state,device_lost gpuweb/cts#882, by @kainino0x.
-
Consolidate helpers for texture content checking gpuweb/cts#881, by @kainino0x.
-
val: render pass state_tracking gpuweb/cts#880, by @kainino0x.
-
val: queue,destroyed,buffer gpuweb/cts#879, by @kainino0x.
-
op: command_buffer,render,state_tracking gpuweb/cts#877, by @Jiawei-Shao.
-
val: copyTextureToTexture 1d textures gpuweb/cts#876, by @kainino0x.
-
op: texture_zero 1d textures gpuweb/cts#875, by @kainino0x.
-
op: copyTextureToTexture 1d textures gpuweb/cts#874, by @kainino0x.
-
Expand tests for 1D textures. gpuweb/cts#872, by @Kangz.
-
test issue for new github projects gpuweb/cts#865, by @kainino0x.
-
7 issues closed.
-
op: command_buffer,render,state_tracking gpuweb/cts#877.
-
val: buffer,create:size gpuweb/cts#936.
-
val: buffer,create:usage gpuweb/cts#937.
-
Consider "subcases" gpuweb/cts#305.
-
Node integration gpuweb/cts#117.
-
test issue for new github projects gpuweb/cts#865.
-
1 issue reopened.
-
val: state,device_lost gpuweb/cts#882, by @kainino0x.
-
-
17 issues received 19 new comments.
-
val: Resource usage validation for textures/buffers gpuweb/cts#905 (2 comments by @Jiawei-Shao, @kainino0x)
-
Update validation,encoding,cmds,copyTextureToTexture:texture_format_equality for srgb compat gpuweb/cts#959 (1 comment by @Kangz)
-
web_platform: context_creation from transferControlToOffscreen gpuweb/cts#908 (1 comment by @kainino0x)
-
web_platform: test copies and imports from videos with color profiles, rotation/crop, etc. gpuweb/cts#910 (1 comment by @kainino0x)
-
How to differentiate between backends due to internal details/rounding gpuweb/cts#930 (1 comment by @zoddicus)
-
-
30 pull requests opened.
-
Adds stress tests for device destroy. gpuweb/cts#961, by @lokokung.
-
Update image copy validation tests for 1D / 3D textures gpuweb/cts#960, by @Kangz.
-
Add cts for copy between non-srgb format texture and srgb format texture gpuweb/cts#956, by @shaoboyan.
-
Add 'flipY' test cases for copyToTexture:* gpuweb/cts#955, by @shaoboyan.
-
Plan api,operation,memory_sync,texture,* gpuweb/cts#954, by @austinEng.
-
Implementing `min` tests gpuweb/cts#953, by @zoddicus.
-
Add rgba8unorm support for swapchain format gpuweb/cts#952, by @jchen10.
-
wgsl: Test that integral user-defined IO requires flat interpolation gpuweb/cts#951, by @jrprice.
-
Build out/{stress,manual}/listing.js so they can be run from online gpuweb/cts#950, by @kainino0x.
-
Factor out numeric case generation code & use it in other tests gpuweb/cts#948, by @zoddicus.
-
wgsl: Add @interpolate(flat) to integral IO gpuweb/cts#947, by @jrprice.
-
Add rgba8unorm support for swapchain format gpuweb/cts#946, by @jchen10.
-
Add rgba8unorm support for swapchain format gpuweb/cts#945, by @jchen10.
-
Add api,operation,command_buffer,render,state_tracking:* -- Part IV gpuweb/cts#944, by @Jiawei-Shao.
-
Add tests for creating buffer size and usage validation gpuweb/cts#943, by @jiangzhaoming.
-
Add rgba8unorm support for swapchain format gpuweb/cts#942, by @jchen10.
-
wgsl: Replace `[[attribute]]` with `@attribute` gpuweb/cts#938, by @ben-clayton.
-
Adds some device.destroy and lost tests. gpuweb/cts#932, by @lokokung.
-
Count TODOs and display them in /standalone/ gpuweb/cts#931, by @kainino0x.
-
Implement `inverseSqrt` tests gpuweb/cts#926, by @zoddicus.
-
Fix api,operation,rendering,blending:GPUBlendComponent:component:* gpuweb/cts#925, by @Jiawei-Shao.
-
Add api,operation,command_buffer,render,state_tracking:* - Part III gpuweb/cts#878, by @Jiawei-Shao.
-
Update createTexture tests with 1D + mipLevelCount > 0 being invalid gpuweb/cts#873, by @Kangz.
-
Documentation updates for new CTS project tracker gpuweb/cts#871, by @kainino0x.
-
Depth-only copy is disallowed for depth24unorm-stencil format gpuweb/cts#870, by @haoxli.
-
Implement `fract` tests gpuweb/cts#869, by @zoddicus.
-
Verify the TODO/DONE state of checklist items gpuweb/cts#868, by @kainino0x.
-
Add api,operation,command_buffer,render,state_tracking:* - Part II gpuweb/cts#867, by @Jiawei-Shao.
-
Validate DepthStencilState in RenderPipeline descriptor gpuweb/cts#866, by @Richard-Yunchao.
-
-
27 pull requests merged.
-
Implementing `min` tests gpuweb/cts#953.
-
Add 'flipY' test cases for copyToTexture:* gpuweb/cts#955.
-
Factor out numeric case generation code & use it in other tests gpuweb/cts#948.
-
Count TODOs and display them in /standalone/ gpuweb/cts#931.
-
Add tests for creating buffer size and usage validation gpuweb/cts#943.
-
Adds some device.destroy and lost tests. gpuweb/cts#932.
-
wgsl: Add @interpolate(flat) to integral IO gpuweb/cts#947.
-
Add rgba8unorm support for swapchain format gpuweb/cts#946.
-
Add rgba8unorm support for swapchain format gpuweb/cts#942.
-
wgsl: Replace `[[attribute]]` with `@attribute` gpuweb/cts#938.
-
Implement `inverseSqrt` tests gpuweb/cts#926.
-
Implement `fract` tests gpuweb/cts#869.
-
Documentation updates for new CTS project tracker gpuweb/cts#871.
-
Depth-only copy is disallowed for depth24unorm-stencil format gpuweb/cts#870.
-
Validate DepthStencilState in RenderPipeline descriptor gpuweb/cts#866.
-
Filter out non-depth formats from depth_test_input_clamped gpuweb/cts#773.
-
Verify the TODO/DONE state of checklist items gpuweb/cts#868.
-
Track all test-coverage TODOs in descriptions gpuweb/cts#864.
-
Improve PR template gpuweb/cts#863.
-
Triage testing tasks gpuweb/cts#862.
-
-
1 pull request closed.
-
Add rgba8unorm support for swapchain format gpuweb/cts#945.
-
-
24 pull requests received 173 new comments.
-
Implementing `min` tests gpuweb/cts#953 (21 comments by @ben-clayton, @dneto0, @kainino0x, @zoddicus)
-
Implement `fract` tests gpuweb/cts#869 (19 comments by @alan-baker, @ben-clayton, @dneto0, @zoddicus)
-
Validate DepthStencilState in RenderPipeline descriptor gpuweb/cts#866 (19 comments by @kainino0x, @Kangz, @Richard-Yunchao)
-
Add tests for creating buffer size and usage validation gpuweb/cts#943 (14 comments by @jiangzhaoming, @kainino0x)
-
Update image copy validation tests for 1D / 3D textures gpuweb/cts#960 (6 comments by @austinEng, @kainino0x, @Kangz)
-
Adds stress tests for device destroy. gpuweb/cts#961 (6 comments by @austinEng, @kainino0x, @lokokung)
-
Factor out numeric case generation code & use it in other tests gpuweb/cts#948 (5 comments by @ben-clayton, @zoddicus)
-
Add api,operation,command_buffer,render,state_tracking:* - Part II gpuweb/cts#867 (5 comments by @Jiawei-Shao, @kainino0x)
-
Add api,operation,command_buffer,render,state_tracking:* -- Part IV gpuweb/cts#944 (4 comments by @Jiawei-Shao, @kainino0x)
-
Add cts for copy between non-srgb format texture and srgb format texture gpuweb/cts#956 (2 comments by @kainino0x, @shaoboyan)
-
Add api,operation,command_buffer,render,state_tracking:* - Part III gpuweb/cts#878 (1 comment by @kainino0x)
-
Update createTexture tests with 1D + mipLevelCount > 0 being invalid gpuweb/cts#873 (1 comment by @lokokung)
-