llama.cpp

e54b3940 - CUDA/HIP: fix ssm_scan on devices where warp size is not 32 (#14196)