Argonne MI300X vLLM Llama-3-8B b64
5244.1 tok/s aggregate
Fields
id
benchmarks/argonne-mi300x-vllm-llama3-8b-fp16-b64-128io-tpsMetric
tps_aggregate
Value
5,244.1
Unit
tok/s aggregate
Model
meta-llama/Meta-Llama-3-8B
Methodology
THIRD-PARTY MEASURED (Argonne LLM-Inference-Bench arXiv:2411.00136). vLLM, FP16/BF16, 128in/128out. throughput=batch*(in+out)/e2e latency = AGGREGATE over batch. 1x bs64 lat 3.92s. CSV 'AMD MI300X GPU' = ALCF 8x MI300X node (xGMI). N = tiles used.
Measured by
third_party
Reproducible
yes
notes
CORRECTED 2026-10-15: value re-read from All_results.csv
Notes
CORRECTED 2026-10-15: value re-read from All_results.csv