HPE · HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)
Llama-2-70B (MLPerf reference) · llama2-70b-99 · Interactive · run 01a00b49-fc74…
valid · Reported evidence
Measurements
output_token_throughput_tps · reported70,954.1 tokens_s
Llama-2-70B (MLPerf reference) · llama2-70b-99 · Interactive · interactive · 8× AMD Instinct MI355X 288GB HBM3e · mlperf_closed_99 · observed 2026-03-26
Identity
modelLlama-2-70B (MLPerf reference)
stackvLLM
workloadllama2-70b-99 · Interactive
topology8× AMD Instinct MI355X 288GB HBM3e
harnessmlperf_loadgen v6.0
quality policymlperf_closed_99
comparison keysha256:0e1f85697886b67d…
sha256:a61c5368fd64020efd08205…
Launch configuration
dtypefp4
notes256 CUs per GPU, gfx950 architecture
systemHPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)
softwarevLLM
availabilityavailable
Workload
namellama2-70b-99 · Interactive
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": "AMD Instinct MI355X 288GB HBM3e",
"acceleratorVendor": "amd",
"acceleratorsPerNode": 8
},
"observedAt": "2026-03-26T19:39:54Z",
"modelDigest": "sha256:8005d6af1339caa02e60709ce2d7a6002ce9c51e436d21bb39dc5b8f93a24b0c",
"measurements": [
{
"unit": "tokens_s",
"value": 70954.1,
"metric": "output_token_throughput_tps",
"statistic": "reported"
}
],
"sourceNative": {
"source": "mlperf-inference",
"metrics": {
"performance_units": "Tokens/s",
"performance_result": 70954.1
},
"category": "datacenter",
"division": "closed",
"scenario": "interactive",
"benchmark": "llama2-70b-99",
"externalId": "v6.0/6.0-0049"
},
"qualityPolicy": "mlperf_closed_99",
"workloadDigest": "sha256:fec425e36d7e44b87a9f9ed1f29e0e9e07a28fb47dd30a4f7a618a7956a6bcf9",
"configurationDigest": "sha256:375395b58450c0efdc019ea4cea6fd4b5cb3b93c3a834c69460b38aba6e662b6"
},
"metadata": {
"name": "mlperf-v6-0-6-0-0049",
"title": "HPE · llama2-70b-99 · Interactive",
"sourceRefs": [
{
"url": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8"
}
]
},
"apiVersion": "kernelindex.dev/v1alpha1"
}Cite this result (permalink, digest, access date)
Report an issue with this serving run
MLPerf Inference · closed/datacenter · entry v6.0/6.0-0049 · shown unmodified as publishedAll serving results →JSON