Greg Roth

Results 67 comments of Greg Roth

#2122 introduced this pass. It seemed targeted for static globals rather than groupshared. That may have been in errror.

Missing validation that is not a regression. It could be masking compiler flaws. If it is, it might be worth getting in with 6.8 as it's an opportunity to make...

I think the first step should be a warning that this doesn't actually do anything. Whether we end up removing it in DXC or not we should discuss.

Solution might involve making the vector template a proper template instead of the hlsl-specific implementation that's there now.

I'll take a quick look to see if anything jumps out. Helana, what kind of time constraints does this have?

It's out! https://github.com/microsoft/DirectXShaderCompiler/releases/tag/v1.8.2405-mesh-nodes-preview

@mheyer32 is the contact on this spec now. These were pretty heavily "inspired" by the desktop specs. It's entirely possible that the enum was inadvertently copied from there.