llama.cpp Vulkan Intel Arc Pro B60 llama-2-7b Q4_0 pp512 FA
274.76 tok/s (pp512)
Fields
id
benchmarks/llamacpp-vulkan-b60-llama27b-q40-fa-pp512Metric
prefill_tok_s
Value
274.76
Unit
tok/s (pp512)
Model
llama-2-7b.Q4_0.gguf (TheBloke, 3.56 GiB, 6.74B params)
Methodology
COMMUNITY-SCOREBOARD (llama.cpp discussion #10879, maintainer-curated Vulkan backend). llama-bench llama-2-7b.Q4_0.gguf (TheBloke, 3.56 GiB, 6.74B params), pp512 + tg128 batch-1, ngl=99 full offload, Vulkan. Build 516a4ca. Chip Intel Arc Pro B60. Per-request batch-1.
Measured by
third_party
Reproducible
yes