Results 545 comments of David Neto

Recent committee discussion https://github.com/gpuweb/gpuweb/issues/5368#issuecomment-3428957336 says: * implement in CTS, and monitor how it behaves across devices * if no surprises, then land the spec change #5416

To fill in some a detail: * First, @kvark is right about the implications of the best-effort framing * That said, the SignedZeroInfNanPreserve was first made available in a Vulkan...

> My reading of the current state of the debate is that we need to decide if this functionality is testable or not. Agreed. I wasn't sure on the call...

Hey @munrocket thanks for this technique! And thank you *also* for a nice compact test case. We had been discussing the need for a good way to test the behaviour....

> That statement would be invalid as the parser is greedy, so it ends up being `(x--) 7`. except that `x--` is a statement, so the `--` token is not...

> I think this shows that it's ambiguous enough to deserve a clarification in the spec :) Yes, definitely we should add an example > a bottom-up parser would have...

Same as #4669 There is more explanation there.

Fixed since subgroup_uniformity extension has landed.

The error bounds on the derivative functions are infinite ULPs. So there's no accuracy issue.