> Authors of CUDA benchmarks used CUDA functions atomicInc and atomicDec which have direct hardware support on NVIDIA cards, but no hardware support on Intel ca

> Authors of CUDA benchmarks used CUDA functions atomicInc and atomicDec which have direct hardware support on NVIDIA cards, but no hardware support on Intel cards. Then how does InterlockedAdd HLSL intrinsic works on Intel? https://docs.microsoft.com/en-us/windows/win32/direct3dhlsl/interlockedadd https://docs.microsoft.com/en-us/windows/win32/direct3dhlsl/...