Intel Arc Pro B70 32GB x4, MLPerf v6.0 Offline: Llama-2-70B aggregate tok/s
2459.18 output tok/s AGGREGATE across 4x Arc Pro B70 (Offline; ~615 tok/s per card)
Fields
id
benchmarks/b70-4x-mlperf-v6-0-llama2-70b-offline-tokensMetric
tps_aggregate
Value
2,459.18
Unit
output tok/s AGGREGATE across 4x Arc Pro B70 (Offline; ~615 tok/s per card)
Model
Llama-2-70B (FP8)
Methodology
VENDOR-MEASURED (MLPerf v6.0 closed submission, Intel). 4x Arc Pro B70 32GB GDDR6 (PCIe Gen5 x16) on Xeon 698X host, PyTorch vLLM, Llama-2-70B, Offline. 'Tokens per second' = 2459.18 verbatim from mlperf_log_summary.txt. AGGREGATE across 4 cards — ~615/card.
Measured by
vendor
Reproducible
yes