6 × NVIDIA B200-SXM-180GB

Tested system

Processor: INTEL(R) XEON(R) PLATINUM 8568Y+
Graphics: NVIDIA B200-SXM-180GB
Accelerator count: 6
Nodes: 1

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

MLPerf Inference / llama2-70b-99.9Offline · v6.1 · Tokens/s

6 × NVIDIA B200-SXM-180GB

130,928^* Tokens/s

Accuracy target / result: ROUGE1: 41.797 ROUGE2: 19.7311 ROUGEL: 26.6864 TOKENS_PER_SAMPLE: 281.9
Weight precision: nvfp4
Division: open
System type: datacenter
Availability: rdi
Cores per processor: 48
Processors per node: 2
Nodes: 1
Total accelerators: 6
Software stack: TensorRT-LLM 1.3.0rc20
Operating system: Ubuntu 24.04.3 LTS (container)

Processing & platform

Processor: INTEL(R) XEON(R) PLATINUM 8568Y+
Graphics: NVIDIA B200-SXM-180GB

Connectivity & software

Operating system: Ubuntu 24.04.3 LTS (container)

System configuration

Reported host core count: 48
Processors per node: 2
Accelerator count: 6
Nodes: 1
Software stack: TensorRT-LLM 1.3.0rc20
Submission division: open

* Provisional data