ASUSTeK · ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8
Llama-2-70B (MLPerf reference) · llama2-70b-99 · Server · run 01a00b49-b17d…
valid · Reported evidence
Measurements
output_token_throughput_tps · reported28,038.5 tokens_s
Llama-2-70B (MLPerf reference) · llama2-70b-99 · Server · server · 8× NVIDIA RTX PRO 6000 Blackwell Server Edition · mlperf_closed_99 · observed 2026-03-26
Identity
modelLlama-2-70B (MLPerf reference)
stackTensorRT 10.14, CUDA 13.1
workloadllama2-70b-99 · Server
topology8× NVIDIA RTX PRO 6000 Blackwell Server Edition
harnessmlperf_loadgen v6.0
quality policymlperf_closed_99
comparison keysha256:504d1ec8161a2333…
sha256:443afaa847e70828ac83737…
Launch configuration
dtypefp4
notesData bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s
systemESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8
softwareTensorRT 10.14, CUDA 13.1
availabilityavailable
Workload
namellama2-70b-99 · 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 RTX PRO 6000 Blackwell Server Edition",
"acceleratorVendor": "nvidia",
"acceleratorsPerNode": 8
},
"observedAt": "2026-03-26T19:39:54Z",
"modelDigest": "sha256:8005d6af1339caa02e60709ce2d7a6002ce9c51e436d21bb39dc5b8f93a24b0c",
"measurements": [
{
"unit": "tokens_s",
"value": 28038.5,
"metric": "output_token_throughput_tps",
"statistic": "reported"
}
],
"sourceNative": {
"source": "mlperf-inference",
"metrics": {
"performance_units": "Tokens/s",
"performance_result": 28038.5
},
"category": "datacenter",
"division": "closed",
"scenario": "server",
"benchmark": "llama2-70b-99",
"externalId": "v6.0/6.0-0004"
},
"qualityPolicy": "mlperf_closed_99",
"workloadDigest": "sha256:ee188c862b2e3bbbca29b194db4aa45d0f33b7da2bdc410aa953d53e6ad6a4ac",
"configurationDigest": "sha256:f5df4a7f8cb966bb5db559dab083da6105cabe83057ed1660dd2616ccf3c72ba"
},
"metadata": {
"name": "mlperf-v6-0-6-0-0004",
"title": "ASUSTeK · llama2-70b-99 · Server",
"sourceRefs": [
{
"url": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT"
}
]
},
"apiVersion": "kernelindex.dev/v1alpha1"
}Cite this result (permalink, digest, access date)
Report an issue with this serving run
MLPerf Inference · closed/datacenter · entry v6.0/6.0-0004 · shown unmodified as publishedAll serving results →JSON