llama.cpp Vulkan on Tesla V100 32GB: Llama-2-7B Q4_0 FA-off tg128 decode tok/s (scoreboard #10879)
129.58 tok/s (tg128, per-request, batch 1)
Fields
id
benchmarks/llamacpp-vulkan-v100-llama27b-q40-nofa-tg128Metric
decode_tok_s
Value
129.58
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 (llama.cpp discussion #10879, maintainer-curated). llama-bench -m llama-2-7b.Q4_0.gguf -ngl 99 -fa 0. Tesla V100 32GB (32 GB HBM2 4096-bit). Build commit 7d77f07, reporter scoreboard. llama-bench default -r 5, mean +/- std. Single-GPU per thread rules. Per-request batch-1, NOT aggregate.
Measured by
third_party
Reproducible
yes
notes
Maintainer-verified scoreboard row from ggml-org/llama.cpp#10879 (Vulkan backend). Non-FA baseline.
Notes
Maintainer-verified scoreboard row from ggml-org/llama.cpp#10879 (Vulkan backend). Non-FA baseline.