KernelTuner/kernel_float
CUDA header-only library for working with vector types (half2, float4, double2) and reduced precision math (half, e5m2) inside kernel code
C++Apache-2.0
CUDA header-only library for working with vector types (half2, float4, double2) and reduced precision math (half, e5m2) inside kernel code
C++Apache-2.0