Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Add shader-float16 extension #1275

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

Add shader-float16 extension #1275

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions spec/index.bs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1351,6 +1351,7 @@ enum GPUFeatureName {
"depth24unorm-stencil8",
"depth32float-stencil8",
"pipeline-statistics-query",
"shader-float16",
"texture-compression-bc",
"timestamp-query",
};
Expand Down Expand Up @@ -7156,6 +7157,13 @@ The following enums are supported if and only if the {{GPUFeatureName/"pipeline-
</dl>


## <dfn dfn-type=enum-value dfn-for=GPUFeatureName>shader-float16</dfn> ## {#shader-float16}

Allows 16bit float arithmetic feature in shader and 16bit storage features for GPUBuffer.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what are the "16bit storage features of GPUBuffer"?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it means accessing buffer memory through 16bit types in WGSL. (but it needs to be reworded to say that)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In principle this is a fine idea.

As noted, there's a lot of detail to be determined. Support in the field is very uneven, and Vulkan slices the functionality into lots of pieces, e.g.

  • some devices can compute in fp16, but not load or store individual fp16 values.

  • some devices can load or store individual fp16 values but not directly do fp16 arithmetic.

  • ability to load/store fp16 values differs by storage class. storage buffer, uniform buffer, input/output, workgroup storage.

  • Also, you'd need to specify all the places f16 can be used, and behaviour of arithmetic, and error bounds on builtins (if they apply)


Issue: link to the 16bit float types, constructors, functions in WGSL spec.


## <dfn dfn-type=enum-value dfn-for=GPUFeatureName>texture-compression-bc</dfn> ## {#texture-compression-bc}

Allows for explicit creation of textures of BC compressed formats.
Expand Down