llama.cpp CUDA H100 80 GB llama-2-7b Q4_0 pp512 noFA
9918.34 tok/s (pp512)
Fields
id
benchmarks/llamacpp-scoreboard-h100-sxm-nofa-pp512Metric
prefill_tok_s
Value
9,918.34
Unit
tok/s (pp512)
Model
llama-2-7b.Q4_0.gguf (ggml-org, 3.56 GiB, 6.74B params)
Methodology
COMMUNITY-SCOREBOARD (llama.cpp discussion #15013, maintainer-curated CUDA backend). llama-bench llama-2-7b.Q4_0.gguf (ggml-org, 3.56 GiB, 6.74B params), pp512 and tg128, n_gpu_layers 99, batch 1, full GPU offload, CUDA. Per-request batch-1, not an aggregate across concurrent requests. Flash attention off (-fa 0). Build 5143fa8, reporter @Hedede. Printed mean 9918.34. Re-read of the discussion page on 2026-10-17.
Measured by
third_party
Reproducible
yes
notes
H100 80 GB, 80 GB / HBM3 / 5120 bit, build 5143fa8, reporter @Hedede. prompt processing cell.
Notes
H100 80 GB, 80 GB / HBM3 / 5120 bit, build 5143fa8, reporter @Hedede. prompt processing cell.