Pavan K Lanka

@rjodinchr one question, given this change went in, is this still accurate (https://github.com/google/clspv/blob/main/docs/OpenCLCOnVulkan.md#built-in-functions) i.e. "This means kernels will operate as if compiled with --cl-fast-relaxed-math" and/or should we need to update...

Thanks @callumfare. Samsung side will pick it up and work with anyone to get this in.

@rjodinchr, Gowtham from Samsung did make progress (which is being internally reviewed) and plan to upload what we got sometime this week. Is that OK or still want to take...

Since, we took it over and/or dont have permissions to codeplaysoftware fork, we went ahead with our own fork PR. And its here. https://github.com/google/clspv/pull/1117 This Pr (https://github.com/google/clspv/pull/925) is referenced there....

@here, excuse me for joining the conversation in the middle. But, trying to connect the dots. IIRC, this test in particular had a bug/issue which I believe was identified/fix back...

Per discussion on 9th Sept OpenCL Call, @kpet pointed out it might be same as https://www.khronos.org/members/login/bugzilla/show_bug.cgi?id=15814 Just documenting that for future reference.