Argonne A100-SXM-40GB x2 llama.cpp mistralai/Mistral-7B-v0.1 b1
68.09 tok/s aggregate (batch*(in+out)/latency)
Fields
id
benchmarks/argonne-a100sxm40-llama-cpp-mistral-7b-fp16-2x-b1-128io-tpsMetric
tps_aggregate
Value
68.09
Unit
tok/s aggregate (batch*(in+out)/latency)
Model
mistralai/Mistral-7B-v0.1
Methodology
THIRD-PARTY MEASURED (Argonne LLM-Inference-Bench, arXiv:2411.00136, All_results.csv). llama.cpp on NVIDIA A100 SXM 40GB x2, FP16/BF16, 128 in / 128 out, batch 1. Throughput = batch*(in+out)/latency = AGGREGATE across batch. Latency 3.76s.
Measured by
third_party
Reproducible
yes
Thermal state
unknown
notes
Argonne hardware matrix: A100 SXM 40GB x4/node. Framework=llama.cpp.
Notes
Argonne hardware matrix: A100 SXM 40GB x4/node. Framework=llama.cpp.