Argonne A100-SXM-40GB x1 llama.cpp meta-llama/Llama-2-7b-hf b64
114.45 tok/s aggregate (batch*(in+out)/latency)
Fields
id
benchmarks/argonne-a100sxm40-llama-cpp-llama2-7b-fp16-1x-b64-128io-tpsMetric
tps_aggregate
Value
114.45
Unit
tok/s aggregate (batch*(in+out)/latency)
Model
meta-llama/Llama-2-7b-hf
Methodology
THIRD-PARTY MEASURED (Argonne LLM-Inference-Bench, arXiv:2411.00136, All_results.csv). llama.cpp on NVIDIA A100 SXM 40GB x1, FP16/BF16, 128 in / 128 out, batch 64. Throughput = batch*(in+out)/latency = AGGREGATE across batch. Latency 2.24s.
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.