llama-bench : add test measuring token generation rate at given prompt length#11126

fairydreaming wants to merge 3 commits intoggml-org:masterggml-org/llama.cpp:masterfrom

fairydreaming:llama-bench-gpfairydreaming/llama.cpp:llama-bench-gpCopy head branch name to clipboard