CUDA: fix race condition in FA vector kernels#13742

JohannesGaessler merged 1 commit intoggml-org:masterggml-org/llama.cpp:masterfrom

JohannesGaessler:cuda-fa-fix-race-cond3JohannesGaessler/llama.cpp:cuda-fa-fix-race-cond3Copy head branch name to clipboard