David Neto

Looks good to me. Thanks

Google has implemented an experimental Chromium extension to help partners progress. It is `chromium-experimental-read-write-storage-texture` and available with extra command line flags on the browser (not meant for end users) We...

> Since we're at it, could we add readonly storage textures? Yes please. That would improve regularity. It removes a weird case: Right now if an author wants a read-only...

> Are there any use cases specifically for rgba32float, rgba32uint, rgba32sint, rgba16float, rgba16uint, rgba16sint, rgba8unorm, rgba8uint, and rgba8sint storage textures that makes them worth the 3rd tier? I don't know...

Ah, I missed that comment. :-) I guess that option would add coupling between the two features; that seems undesirable? I lean toward not coupling features.

> The table above makes me wonder whether rgba32float, rgba32uint, rgba32sint, rgba16float, rgba16uint, rgba16sint, rgba8unorm, rgba8uint, and rgba8sint on unextended Vulkan are worth the 3rd tier level. It would totally...

Another thing that is a gotcha is that the spec usually talks about vector types in the fully-elaborated template spelling, e.g. `vec3` rather than the shorter alias `vec3f`. Always look...

FYI. @gfxstrand blogged about the what's needed to enforce "maximal reconvergence" on NV hardware. It's quite challenging, and it definitely doesn't happen by accident. ~https://www.khronos.org/blog/khronos-releases-maximal-reconvergence-and-quad-control-extensions-for-vulkan-and-spir-v~ ---- edit: Whoops should have...

The proposal talks a bit about non-full subgroups (when the subgroup size does not evenly divide the workgroup size): > TODO: Can we add a pipeline parameter to require full...

Should there be shader-creation-time (or pipeline-creation-time) checks on certain parameters when they are const (or override, respectively): * subgroupShuffleXor mask operand: if it's const it must be < 128 (which...