llama.cpp Vulkan on Intel Arc Pro B70: Llama-2-7B Q4_0 noFA tg128 decode tok/s
112.02 tok/s (tg128, per-request, batch 1)
Fields
id
benchmarks/llamacpp-vulkan-arc-pro-b70-llama27b-q40-nofa-tg128Metric
decode_tok_s
Value
112.02
Unit
tok/s (tg128, per-request, batch 1)
Model
llama-2-7b.Q4_0.gguf (TheBloke, 3.56 GiB, 6.74B params)
Methodology
COMMUNITY-SCOREBOARD (#10879). llama-bench -ngl 100 -fa 0, Vulkan backend. Build b863507. Per-request batch-1.
Measured by
third_party
Reproducible
yes
notes
Vulkan decode on B70 (112) is competitive with ROCm/CUDA mid-range cards on the same workload.
Notes
Vulkan decode on B70 (112) is competitive with ROCm/CUDA mid-range cards on the same workload.