CUDA: missing PDL sync for FWHT, better fallback#23690

am17an merged 1 commit into

ggml-org:masterggml-org/llama.cpp:masterfrom

JohannesGaessler:cuda-fix-fwht-pdlJohannesGaessler/llama.cpp:cuda-fix-fwht-pdlCopy head branch name to clipboard