GPU for the Web Working Group - Activity
No matches.
Participants
-
2 participants joined.
-
Pratik Ashok Thakare (ByteDance) joined.
-
Romaric JODIN (Google LLC) joined.
-
-
1 participant left.
-
Anders Leino (NVIDIA) left.
-
GitHub
gpuweb/cts
-
1 issue closed.
-
Add a sampling operation test for `texture-compression-unaligned`? gpuweb/cts#4707.
-
-
2 issues received 6 new comments.
-
Add a sampling operation test for `texture-compression-unaligned`? gpuweb/cts#4707 (4 comments by @castano, @Kangz)
-
shader/execution/shader_io/fragment_builtins expects more interpolation precision than specs guarantee gpuweb/cts#4657 (2 comments by @dneto0, @fooishbar)
-
-
3 pull requests opened.
-
Validate sampling operations on unaligned block-compressed textures. gpuweb/cts#4717, by @castano.
-
Fix configure:invalid_usage for bgra8unorm storage gpuweb/cts#4716, by @sudonatalie.
-
Add more tests for mixing up depth and color attachments gpuweb/cts#4715, by @kainino0x.
-
-
2 pull requests merged.
-
Fix configure:invalid_usage for bgra8unorm storage gpuweb/cts#4716.
-
Validate sampling operations on unaligned block-compressed textures. gpuweb/cts#4717.
-
-
3 pull requests received 7 new comments.
-
Validate sampling operations on unaligned block-compressed textures. gpuweb/cts#4717 (4 comments by @castano, @Kangz)
-
Add unary, function, and uniformity tests for WGSL immediate pointers gpuweb/cts#4677 (1 comment by @shaoboyan091)
-
gpuweb/gpuweb
-
2 issues opened.
-
Consider optional FP8 operand formats for subgroup matrices gpuweb/gpuweb#11051, by @sayanshaw24.
-
[Time sensitive] Rename WGSL feature "fragment_depth" gpuweb/gpuweb#11050, by @AshleyScirra.
-
-
3 issues received 3 new comments.
-
WGSL Proposal for fragment depth (less, greater, any) gpuweb/gpuweb#5342 (1 comment by @AshleyScirra)
-
wgsl: support atomics on f32 (in workgroup and storage buffer) gpuweb/gpuweb#4894 (1 comment by @akre54)
-
-
2 pull requests opened.
-
Update filterability document gpuweb/gpuweb#11049, by @dj2.
-
Fix bindless typos gpuweb/gpuweb#11048, by @dj2.
-
-
2 pull requests merged.
-
Update filterability document gpuweb/gpuweb#11049.
-
Fix bindless typos gpuweb/gpuweb#11048.
-
-
4 pull requests received 5 new comments.
-
Update validity conditions of device & buffer on mapAsync gpuweb/gpuweb#10971 (1 comment by @sagudev)
TR publications
-
4 documents published.
-
WebGPU Shading Language published twice as a Candidate Recommendation Draft. History of WebGPU Shading Language
-
WebGPU published twice as a Candidate Recommendation Draft. History of WebGPU
-
Participants
-
1 participant joined.
-
Felix Krall (NVIDIA) joined.
-
-
4 participants left.
-
Daniel Crookston (Autodesk) left.
-
Emre Yalcinoglu (Autodesk) left.
-
Aura Munoz (Autodesk) left.
-
Henrik Edstrom (Autodesk) left.
-
-
1 organization left.
-
Autodesk left.
-
GitHub
gpuweb/gpuweb
-
3 issues opened.
-
maxComputeWorkgroupsPerDimension should be configurable per dimension gpuweb/gpuweb#10963, by @MonaMayrhofer.
-
Adapter state is never set to "consumed" gpuweb/gpuweb#10469, by @sagudev.
-
Persistent Buffer Mapping gpuweb/gpuweb#7290, by @kainino0x.
-
-
8 issues closed.
-
Coarse "available memory" signal on GPUAdapter for capacity planning gpuweb/gpuweb#6957.
-
JPEG XL and WebGPU gpuweb/gpuweb#6879.
-
Make ETC2 and ASTC always allow unaligned sizes? gpuweb/gpuweb#6956.
-
[bindless] Add aliasing updates for bindless to WGSL spec gpuweb/gpuweb#5581.
-
Add limited support for 64 bit atomics gpuweb/gpuweb#5071.
-
[bindless] Define the default types gpuweb/gpuweb#5471.
-
[bindless] Define the compatibiilty rules for WGSL types gpuweb/gpuweb#5470.
-
-
1 issue reopened.
-
Adapter state is never set to "consumed" gpuweb/gpuweb#10469, by @kainino0x.
-
-
22 issues received 43 new comments.
-
proposal: eliminate the aliasing restrictions gpuweb/gpuweb#5611 (9 comments by @alan-baker, @jimblandy, @mwyrzykowski)
-
maxComputeWorkgroupsPerDimension should be configurable per dimension gpuweb/gpuweb#10963 (3 comments by @kainino0x, @stefnotch)
-
Metal does not support some texture views of transient attachments that WebGPU allows gpuweb/gpuweb#6876 (3 comments by @kainino0x, @mwyrzykowski)
-
[bindless] Add aliasing updates for bindless to WGSL spec gpuweb/gpuweb#5581 (2 comments by @jimblandy)
-
Are "Fine" derivatives guaranteed to actually be fine? gpuweb/gpuweb#4325 (2 comments by @dneto0, @jimblandy)
-
Coarse "available memory" signal on GPUAdapter for capacity planning gpuweb/gpuweb#6957 (1 comment by @kainino0x)
-
Let developers statically assert uniformity of control flow and expressions. gpuweb/gpuweb#6955 (1 comment by @jimblandy)
-
Lift requirement that alphaToCoverageEnabled attachment have an alpha channel gpuweb/gpuweb#5562 (1 comment by @kainino0x)
-
Should alphaToCoverageEnabled require a blendable format? gpuweb/gpuweb#4506 (1 comment by @kainino0x)
-
Support "uniform texel buffer" and "storage texel buffer" ? gpuweb/gpuweb#162 (1 comment by @jimblandy)
-
wgsl: add feature to enable more IEEE 754 floating point behaviours gpuweb/gpuweb#2259 (1 comment by @savant117)
-
Why do block-compressed textures have to be aligned to a multiple of the texel block size? gpuweb/gpuweb#2006 (1 comment by @kainino0x)
-
-
11 pull requests opened.
-
Clarify immediate variables and fix immediate-data examples gpuweb/gpuweb#11047, by @akaashrp.
-
Update validity conditions of device & buffer on mapAsync gpuweb/gpuweb#10971, by @sagudev.
-
[subgroup-matrix] Update proposal for D3D12 gpuweb/gpuweb#10964, by @alan-baker.
-
Check new preview system 20260922 gpuweb/gpuweb#10962, by @mehmetoguzderin.
-
Serve PR previews through the GPUWeb Pages viewer gpuweb/gpuweb#10961, by @mehmetoguzderin.
-
[bindless] Allow non-uniform indexing gpuweb/gpuweb#10960, by @alan-baker.
-
Add a proposal for uniformity assertions gpuweb/gpuweb#10959, by @alan-baker.
-
[editorial] Move atomic-vec2u-min-max to chronological position gpuweb/gpuweb#10497, by @kainino0x.
-
Set adapter.state to "consumed" gpuweb/gpuweb#10470, by @sagudev.
-
[bindless] Update proposal for WGSL aliasing extension. gpuweb/gpuweb#10345, by @dj2.
-
[wgsl] Add proposal for WGSL namespaces. gpuweb/gpuweb#7310, by @dj2.
-
-
6 pull requests merged.
-
Set adapter.state to "consumed" gpuweb/gpuweb#10470.
-
Serve PR previews through the GPUWeb Pages viewer gpuweb/gpuweb#10961.
-
[bindless] Update proposal for WGSL aliasing extension. gpuweb/gpuweb#10345.
-
[wgsl] Add proposal for WGSL namespaces. gpuweb/gpuweb#7310.
-
Spec changes for atomic vec2u gpuweb/gpuweb#5610.
-
-
1 pull request closed.
-
Check new preview system 20260922 gpuweb/gpuweb#10962.
-
-
8 pull requests received 36 new comments.
-
[wgsl] Add proposal for WGSL namespaces. gpuweb/gpuweb#7310 (17 comments by @BenjaminBrienen, @dj2, @jimblandy, @mehmetoguzderin, @mighdoll, @stefnotch)
-
Add a proposal for uniformity assertions gpuweb/gpuweb#10959 (4 comments by @alan-baker, @Kangz, @mehmetoguzderin)
-
[subgroup-matrix] Update proposal for D3D12 gpuweb/gpuweb#10964 (2 comments by @alan-baker, @amaiorano)
-
gpuweb/cts
-
4 issues opened.
-
CTS expects wrong result of mapAsync on invalid buffer gpuweb/cts#4708, by @sagudev.
-
Add a sampling operation test for `texture-compression-unaligned`? gpuweb/cts#4707, by @Kangz.
-
Add createBindGroupLayout test for exactly one binding type gpuweb/cts#4704, by @sagudev.
-
-
1 issue closed.
-
Add createBindGroupLayout test for exactly one binding type gpuweb/cts#4704.
-
-
4 issues received 11 new comments.
-
CTS expects wrong result of mapAsync on invalid buffer gpuweb/cts#4708 (6 comments by @kainino0x, @sagudev)
-
Add a sampling operation test for `texture-compression-unaligned`? gpuweb/cts#4707 (2 comments by @castano, @Kangz)
-
refactor `webgpu:api,operation,command_buffer,queries,timestampQuery:many_query_sets` gpuweb/cts#4700 (2 comments by @kainino0x)
-
Add createBindGroupLayout test for exactly one binding type gpuweb/cts#4704 (1 comment by @kainino0x)
-
-
12 pull requests opened.
-
Pointer aliasing gpuweb/cts#4714, by @alan-baker.
-
mapAsync on destroyed device gpuweb/cts#4713, by @sagudev.
-
Buffer view tests for local aliasing gpuweb/cts#4712, by @alan-baker.
-
Fix stale test description in the requestDevice:stale: test gpuweb/cts#4711, by @kainino0x.
-
Remove 'as GPUFeatureName' casts that are no longer needed. gpuweb/cts#4710, by @Kangz.
-
Test alphaToCoverageEnabled constraints gpuweb/cts#4706, by @Kangz.
-
Add `createBindGroupLayout` test for exactly one binding type gpuweb/cts#4705, by @sagudev.
-
Add vector assignment execution tests gpuweb/cts#4703, by @sudonatalie.
-
Add a test that requires all known WGSL language features to be supported gpuweb/cts#4702, by @kainino0x.
-
Add lint to ensure kKnownWGSLLanguageFeatures is complete gpuweb/cts#4701, by @kainino0x.
-
Add chained swizzle tests gpuweb/cts#4699, by @sudonatalie.
-
webgpu: Cover immediate data in indirect commands gpuweb/cts#4698, by @shaoboyan091.
-
-
10 pull requests merged.
-
Buffer view tests for local aliasing gpuweb/cts#4712.
-
Fix stale test description in the requestDevice:stale: test gpuweb/cts#4711.
-
Remove 'as GPUFeatureName' casts that are no longer needed. gpuweb/cts#4710.
-
Test alphaToCoverageEnabled constraints gpuweb/cts#4706.
-
Add `createBindGroupLayout` test for exactly one binding type gpuweb/cts#4705.
-
Add vector assignment execution tests gpuweb/cts#4703.
-
Add lint to ensure kKnownWGSLLanguageFeatures is complete gpuweb/cts#4701.
-
Add chained swizzle tests gpuweb/cts#4699.
-
Allow 1 ULP on canvas import tests as noted in the comment. gpuweb/cts#4697.
-
-
2 pull requests received 7 new comments.
TR publications
-
10 documents published.
-
WebGPU Shading Language published 7 times as a Candidate Recommendation Draft. History of WebGPU Shading Language
-
WebGPU published 3 times as a Candidate Recommendation Draft. History of WebGPU
-
Participants
-
1 Team contact appointed.
-
Bert Bos became a Team contact.
-
-
1 participant joined.
-
Bert Bos (W3C) joined.
-
GitHub
gpuweb/gpuweb
-
12 issues opened.
-
Streaming network resources into GPU resources gpuweb/gpuweb#6968, by @kainino0x.
-
Coarse "available memory" signal on GPUAdapter for capacity planning gpuweb/gpuweb#6957, by @hemanth.
-
Make ETC2 and ASTC always allow unaligned sizes? gpuweb/gpuweb#6956, by @kainino0x.
-
Let developers statically assert uniformity of control flow and expressions. gpuweb/gpuweb#6955, by @Kangz.
-
Swizzle View spec issues gpuweb/gpuweb#6941, by @BenjaminBrienen.
-
JPEG XL and WebGPU gpuweb/gpuweb#6879, by @greggman.
-
Metal does not support some texture views of transient attachments that WebGPU allows gpuweb/gpuweb#6876, by @andyleiserson.
-
Depth format table should call out optional features gpuweb/gpuweb#6368, by @kainino0x.
-
WGSL Bindless Filterability Validation gpuweb/gpuweb#6364, by @jrprice.
-
Cannot reliably specify depth biases with depth24plus gpuweb/gpuweb#6350, by @magcius.
-
WGSL vs API dependent feature requirements gpuweb/gpuweb#6347, by @alan-baker.
-
Disambiguate "T is concrete numeric scalar or numeric vector" gpuweb/gpuweb#6346, by @k2d222.
-
-
17 issues closed.
-
Swizzle View spec issues gpuweb/gpuweb#6941.
-
[bindless] Add the proposal for how bindings are updated in resource tables. gpuweb/gpuweb#5379.
-
Clarifications around memory locations gpuweb/gpuweb#3029.
-
Assignment of swizzled values should be permitted gpuweb/gpuweb#737.
-
[bindless] Add the proposed WGSL additions gpuweb/gpuweb#5380.
-
Reconsider "snorm10-10-10-2" vertex format support? gpuweb/gpuweb#6292.
-
WGSL Proposal for fragment depth (less, greater, any) gpuweb/gpuweb#5342.
-
WGSL vs API dependent feature requirements gpuweb/gpuweb#6347.
-
[buffer-view] Extent of reinterpretation permitted gpuweb/gpuweb#5409.
-
[buffer-view] Settle on minimum binding size requirements gpuweb/gpuweb#5410.
-
[buffer-view] Allowed size expressions for buffers gpuweb/gpuweb#5486.
-
subgroup-id proposal discussions gpuweb/gpuweb#5365.
-
Applying `blend_src` to function parameters and return types? gpuweb/gpuweb#5552.
-
Disambiguate "T is concrete numeric scalar or numeric vector" gpuweb/gpuweb#6346.
-
-
32 issues received 70 new comments.
-
Metal does not support some texture views of transient attachments that WebGPU allows gpuweb/gpuweb#6876 (7 comments by @andyleiserson, @kainino0x, @matthargett)
-
Cannot reliably specify depth biases with depth24plus gpuweb/gpuweb#6350 (6 comments by @jimblandy, @kainino0x, @Kangz, @magcius)
-
Feature detection for new core API capabilities gpuweb/gpuweb#6337 (6 comments by @beaufortfrancois, @kainino0x, @Kangz, @mwyrzykowski)
-
Reconsider "snorm10-10-10-2" vertex format support? gpuweb/gpuweb#6292 (4 comments by @jimblandy, @Kangz, @wangra-google)
-
Coarse "available memory" signal on GPUAdapter for capacity planning gpuweb/gpuweb#6957 (3 comments by @hemanth, @kainino0x)
-
[bindless] Add uniformity requirements for bind-less to WGSL spec gpuweb/gpuweb#5582 (3 comments by @alan-baker, @jimblandy, @Kangz)
-
WGSL vs API dependent feature requirements gpuweb/gpuweb#6347 (3 comments by @jimblandy, @mwyrzykowski)
-
Let developers statically assert uniformity of control flow and expressions. gpuweb/gpuweb#6955 (2 comments by @dneto0)
-
function workgroup-memory args should be more flexible (WGSL proposed language change) gpuweb/gpuweb#5156 (2 comments by @jimblandy)
-
Add method to disable fast-math on a per-shader basis. gpuweb/gpuweb#2076 (2 comments by @beaufortfrancois, @isghe)
-
[Wiki] Implementation Status: Firefox Android about:config and Chrome Android status gpuweb/gpuweb#5521 (2 comments by @kainino0x, @VariableVince)
-
Dynamically choosing a render target array index slice from the vertex shader gpuweb/gpuweb#4812 (1 comment by @Kangz)
-
allow declaring workgroup memory within function scope (WGSL proposed language change) gpuweb/gpuweb#5155 (1 comment by @jimblandy)
-
builtin variables should be automatically available inside kernels (WGSL proposed language change) gpuweb/gpuweb#5153 (1 comment by @jimblandy)
-
[buffer-view] Settle on minimum binding size requirements gpuweb/gpuweb#5410 (1 comment by @alan-baker)
-
Applying `blend_src` to function parameters and return types? gpuweb/gpuweb#5552 (1 comment by @dneto0)
-
Disambiguate "T is concrete numeric scalar or numeric vector" gpuweb/gpuweb#6346 (1 comment by @dneto0)
-
Can fragment output location be of `vecN<f16>` type? gpuweb/gpuweb#5551 (1 comment by @andyleiserson)
-
-
12 pull requests opened.
-
spec/index.bs: add local biblio entry for WGSL gpuweb/gpuweb#7285, by @dneto0.
-
wgsl: swizzle view support indexing: delete stale note gpuweb/gpuweb#7065, by @dneto0.
-
wgsl: convert markdown blockquote to explicit HTML gpuweb/gpuweb#6878, by @dneto0.
-
[wgsl] Add a note about alignments gpuweb/gpuweb#6412, by @alan-baker.
-
[wgsl] Auto-enable dependent features gpuweb/gpuweb#6366, by @alan-baker.
-
[subgroup-matrix] Add stride restriction gpuweb/gpuweb#6365, by @alan-baker.
-
[editorial] Editing on Feature Index gpuweb/gpuweb#6351, by @kainino0x.
-
[wgsl]: clarify where attributes can be used on function params and r… gpuweb/gpuweb#6349, by @dneto0.
-
clarify "T is concrete numeric scalar or concrete numeric vector" gpuweb/gpuweb#6348, by @dneto0.
-
[subgroup-matrix] Relax load/store array requirements gpuweb/gpuweb#6345, by @alan-baker.
-
[editorial/bugfix] Dedup feature dependencies, fix missing auto-enablements gpuweb/gpuweb#6344, by @kainino0x.
-
-
13 pull requests merged.
-
wgsl: swizzle view support indexing: delete stale note gpuweb/gpuweb#7065.
-
wgsl: support swizzle assignment gpuweb/gpuweb#5268.
-
wgsl: convert markdown blockquote to explicit HTML gpuweb/gpuweb#6878.
-
[wgsl] Integrate the fragment_depth proposal into the spec gpuweb/gpuweb#6299.
-
[wgsl] Add a note about alignments gpuweb/gpuweb#6412.
-
[wgsl] Auto-enable dependent features gpuweb/gpuweb#6366.
-
[subgroup-matrix] Add stride restriction gpuweb/gpuweb#6365.
-
[editorial] Editing on Feature Index gpuweb/gpuweb#6351.
-
[editorial/bugfix] Dedup feature dependencies, fix missing auto-enablements gpuweb/gpuweb#6344.
-
[wgsl]: clarify where attributes can be used on function params and r… gpuweb/gpuweb#6349.
-
clarify "T is concrete numeric scalar or concrete numeric vector" gpuweb/gpuweb#6348.
-
[subgroup-matrix] Relax load/store array requirements gpuweb/gpuweb#6345.
-
-
13 pull requests received 66 new comments.
-
Add HTML-in-Canvas proposal gpuweb/gpuweb#6250 (23 comments by @kainino0x, @mwyrzykowski, @szager-chromium)
-
[subgroup-matrix] Relax load/store array requirements gpuweb/gpuweb#6345 (6 comments by @dneto0, @jrprice)
-
Add `texture-compression-unaligned` feature to relax texture size block alignment requirement. gpuweb/gpuweb#6312 (2 comments by @kainino0x, @Kangz)
-
[editorial/bugfix] Dedup feature dependencies, fix missing auto-enablements gpuweb/gpuweb#6344 (2 comments by @toji)
-
Adds the `snorm10-10-10-2` vertex format to the WebGPU specification and WebIDL definition. gpuweb/gpuweb#6491 (1 comment by @kainino0x)
-
[wgsl] Integrate the fragment_depth proposal into the spec gpuweb/gpuweb#6299 (1 comment by @dneto0)
-
gpuweb/cts
-
3 issues opened.
-
Canvas import tests can vary by 1 ULP gpuweb/cts#4696, by @mwyrzykowski.
-
Error 404 on cts page gpuweb/cts#4691, by @BenjaminBrienen.
-
-
1 issue closed.
-
Error 404 on cts page gpuweb/cts#4691.
-
-
1 issue received 1 new comment.
-
5 pull requests opened.
-
Allow 1 ULP on canvas import tests as noted in the comment. gpuweb/cts#4697, by @mwyrzykowski.
-
Adds CTS test coverage for the `snorm10-10-10-2` vertex format. gpuweb/cts#4695, by @wangra-google.
-
Subgroup_size_control enables subgroups gpuweb/cts#4693, by @alan-baker.
-
Make README.md links work on github.io gpuweb/cts#4692, by @kainino0x.
-
webgpu: speed up sample_texture_combos gpuweb/cts#4690, by @haoxli.
-
-
5 pull requests merged.
-
Adds CTS test coverage for the `snorm10-10-10-2` vertex format. gpuweb/cts#4695.
-
Subgroup_size_control enables subgroups gpuweb/cts#4693.
-
Add frag_depth qualifier execution test gpuweb/cts#4689.
-
Make README.md links work on github.io gpuweb/cts#4692.
-
webgpu: speed up sample_texture_combos gpuweb/cts#4690.
-
-
1 pull request closed.
-
5 pull requests received 12 new comments.
-
Adds CTS test coverage for the `snorm10-10-10-2` vertex format. gpuweb/cts#4695 (8 comments by @kainino0x, @wangra-google)
-
Add unary, function, and uniformity tests for WGSL immediate pointers gpuweb/cts#4677 (1 comment by @jrprice)
-