Stuart Brady comments

Results 5 comments of


                                            Stuart Brady

object in atomic_load should be declared as const

If we applied this as a bug fix, it would mean that existing implementations that lack the const qualifier would not conform to the specification. No existing working OpenCL C...

SPIR-V Environment Spec does not describe required accuracy for OpFMod and OpFRem

Could you help explain a little about the layering with Vulkan, please? Naïvely, I had previously assumed the translation of _OpFRem_ instructions in SPIR-V using the Kernel execution into SPIR-V...

SPIR-V Environment Spec does not describe required accuracy for OpFMod and OpFRem

As noted in today's WG call, _FPFastMathMode_ decorations are always valid for _OpFRem_, but they are only valid for _fmod_ in the _OpenCL Extended Instruction Set_ when using a SPIR-V...

SPIR-V Environment Spec does not describe required accuracy for OpFMod and OpFRem

I've been asked to clarify my concerns with defining the required accuracy of _OpFRem_ and _OpFMod_ in terms of sequences using _trunc_ and _floor_. I don't have any specific cases...

Missing Spec Descriptions for CL_UNORM_SHORT_565 and CL_UNORM_SHORT_555

@b-sumner, that seems to be defined in the OpenCL 2.x specifications § 5.3.1.1 _Image Format Descriptor_, as follows: > For CL_UNORM_SHORT_565, `R` is in bits `15:11`, `G` is in bits...