llama.cpp Vulkan Intel Arc A750 llama-2-7b Q4_0 tg128 FA
43.96 tok/s (tg128)
Fields
id
benchmarks/llamacpp-vulkan-arc-a750-llama27b-q40-fa-tg128Metric
decode_tok_s
Value
43.96
Unit
tok/s (tg128)
Model
llama-2-7b.Q4_0.gguf (ggml-org, 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 (ggml-org, 3.56 GiB, 6.74B params), pp512 and tg128, batch 1, full GPU offload, Vulkan. Per-request batch-1, not aggregate across concurrent requests. Flash attention on (-fa 1). Build c1b1876. Chip Intel Arc A750. Printed mean 43.96 +/- 0.03. Generating 128 tokens, batch 1.
Measured by
third_party
Reproducible
yes
notes
Flash-attention scoreboard row. Decode is a small gain over noFA (43.96 vs 42.66).
Notes
Flash-attention scoreboard row. Decode is a small gain over noFA (43.96 vs 42.66).