vulkan: tune mat-vec rows for batched inference on Strix Halo - #27909

vulkan: tune mat-vec rows for batched inference on Strix Halo#27909

ggerganov merged 2 commits into

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

SimonTeixidor:strix-halo-vulkan-tuningSimonTeixidor/llama.cpp:strix-halo-vulkan-tuningCopy head branch name to clipboard