Closed apcameron closed 1 year ago
Unfortunately, we have currently no plans for adding non-CUDA GPU kernels. However, I suspect that, in principle, similar techniques as the ones we are using for our CUDA kernels probably also work in other GPU / accelerator programming frameworks, hence quant_cuda.cpp
may be a helpful (higher-level) reference for such implementations.
Please add OpenCL Support that so that it can be used on GPU's that Support OpenCL and not CUDA Then we could use something like quant_opencl.cpp instead of quant_cuda.cpp