NVIDIA · NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)
Llama-3.1-8B (MLPerf reference) · llama3.1-8b · Server · run 01a00b4a-2570…
valid · Reported evidence
Measurements
output_token_throughput_tps · reported141,565 tokens_s
Llama-3.1-8B (MLPerf reference) · llama3.1-8b · Server · server · 8× NVIDIA B300-SXM-270GB · mlperf_closed_99 · observed 2026-03-26
Identity
modelLlama-3.1-8B (MLPerf reference)
stackTensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentM
workloadllama3.1-8b · Server
topology8× NVIDIA B300-SXM-270GB
harnessmlperf_loadgen v6.0
quality policymlperf_closed_99
comparison keysha256:04c21b6522654a17…
sha256:49a4f37551c2c84234fa257…
Launch configuration
dtypefp4
notesB300 TGP 1100W
systemNVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)
softwareTensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0
availabilityavailable
Workload
namellama3.1-8b · Server
streamingyes
load generationopen_loop
Canonical manifest
Show manifestHide manifest
{
"kind": "ServingRun",
"spec": {
"status": "valid",
"harness": {
"name": "mlperf_loadgen",
"version": "v6.0"
},
"topology": {
"nodeCount": 1,
"acceleratorModel": "NVIDIA B300-SXM-270GB",
"acceleratorVendor": "nvidia",
"acceleratorsPerNode": 8
},
"observedAt": "2026-03-26T19:39:54Z",
"modelDigest": "sha256:dafdb8bb411a12b698d36856eed69f16242fb467999fb68680ca09a5601b6bd4",
"measurements": [
{
"unit": "tokens_s",
"value": 141565,
"metric": "output_token_throughput_tps",
"statistic": "reported"
}
],
"sourceNative": {
"source": "mlperf-inference",
"metrics": {
"performance_units": "Tokens/s",
"performance_result": 141565
},
"category": "datacenter",
"division": "closed",
"scenario": "server",
"benchmark": "llama3.1-8b",
"externalId": "v6.0/6.0-0073"
},
"qualityPolicy": "mlperf_closed_99",
"workloadDigest": "sha256:209ab3ba41ef397171c081377f84bcb2512e3d47cd6334278a712a3858d8eedd",
"configurationDigest": "sha256:3084ddd2dfe5499ce5ea3895e30edceea6b82dbd2febad0ee4d4efa55da11c6e"
},
"metadata": {
"name": "mlperf-v6-0-6-0-0073",
"title": "NVIDIA · llama3.1-8b · Server",
"sourceRefs": [
{
"url": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT"
}
]
},
"apiVersion": "kernelindex.dev/v1alpha1"
}Cite this result (permalink, digest, access date)
Report an issue with this serving run
MLPerf Inference · closed/datacenter · entry v6.0/6.0-0073 · shown unmodified as publishedAll serving results →JSON