MLPerf v6.0 Intel 4x Intel Arc Pro B70: llama2-70b-99.9 Offline aggregate tok/s
2459.18 output tokens/s (aggregate over 4 accelerators, Offline)
Fields
id
benchmarks/mlperf-v6-0-intel-4xarcb70-llama2-70b-llama2-70b-99-9-offlineMetric
tps_aggregate
Value
2,459.18
Unit
output tokens/s (aggregate over 4 accelerators, Offline)
Model
llama2-70b-99.9
Methodology
VENDOR-SUBMITTED MLPerf Inference v6.0 closed division, Datacenter suite, Offline scenario. Intel 1-node-4x-BMG-B70: 4x Intel Arc Pro B70. Framework PyTorch vLLM. 'Tokens/s' = AGGREGATE output-token throughput across the 4 accelerators (NOT per-request). Value verbatim from summary_results.json. min_query_count/min_duration per MLPerf rules; binding latency constraint applies to Server/Interactive.
Measured by
vendor
Reproducible
yes