deepgemm_fp8_fp4_gemm_nt_n4096_k4096
GEMM FP8 fp4 n4096 k4096 · m = 1024 · fp32/fp8_e4m3 · run 01a00b1b-8634…
●passedReported evidence
Primary measurement
26.4µsmean
Rank 1 in its comparison group · source-native comparison · observed 2026-06-06
43.89× reference
Identity
implementationdeepgemm_fp8_fp4_gemm_nt_n4096_k4096
projectFlashInfer-Bench baselines
revisionda915083d4c7c5e61aa3005e3d17ae488e0fc71c
workloadm = 1024 · fp32/fp8_e4m3
comparison keysha256:db211fcdfa975179…
sourceFlashInfer-Bench
external idtrace/deepgemm_fp8_fp4_gemm_nt_n4096_k4096/32bb0cb6-daa5-4c2d-aee7-036ae5482287
sha256:d20d53062c4247c5e490940…
Correctness
comparatorflashinfer_bench_eval
max absolute error1
max relative error0.03571261465549469
resultpassed
Workload
m1024
a_fp8fp8_e4m3 [1024, 4096]
b_fp4int8 [4096, 2048]
a_scalefp32 [1024, 32]
b_scalefp32 [4096, 128]
definition comparatorsol_execbench_eval
Measurements
latency · mean26.4 µs
Protocol
harnessflashinfer-bench
timerharness_reported
primaryStatisticmean
correctnessReferencedefinition reference implementation
comparabilityFamilyflashinfer_bench_trace
Environment
gpuNVIDIA B200 (unknown)
cudaToolkit13.0
frameworkpytorch 2.11.0+cu130
Artifacts
No artifacts published with this run.
Replications and notes
No attestations yet.
Community attestations. They never change the evidence level; only a KernelIndex-controlled rerun does.
Add a reproduction or note
Canonical manifest
Show manifestHide manifest
{
"run": {
"kind": "BenchmarkRun",
"spec": {
"status": "passed",
"timing": {
"latencyNs": {
"mean": 26416
},
"primaryStatistic": "mean"
},
"observedAt": "2026-06-06T02:49:24.357Z",
"correctness": {
"comparator": "sol_execbench_eval",
"maximumAbsoluteError": 1,
"maximumRelativeError": 0.03571261465549469
},
"sourceNative": {
"source": "flashinfer-bench",
"metrics": {
"latency_ms": 0.026416,
"speedup_factor": 43.89426862507571,
"reference_latency_ms": 1.159511
},
"benchmark": "gemm_fp8_fp4_n4096_k4096"
},
"protocolDigest": "sha256:7d44d206ad086b4bece35ce31f6531ebb11f07e0d9dcfb74c75be86d3ec02095",
"workloadDigest": "sha256:12f270e05bf6d871edbfac761229c416b5e607fa1249842bcb002f0fb5443011",
"environmentDigest": "sha256:83c9b6e78d493927f18878d57448d9a0d2221a70840b562455b2665a2897ba19",
"implementationDigest": "sha256:34e8889aa8c54757cf304da2d43a71dbe8cfef8c6064ec1ec59448b74be7d1a1"
},
"metadata": {
"name": "sol-trace-gemm-fp8-fp4-n4096-k4096-32bb0cb6",
"title": "deepgemm_fp8_fp4_gemm_nt_n4096_k4096 · gemm_fp8_fp4_n4096_k4096"
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"protocol": {
"kind": "BenchmarkProtocol",
"spec": {
"harness": {
"name": "flashinfer-bench",
"repository": "https://github.com/flashinfer-ai/flashinfer-bench"
},
"correctness": {
"reference": "definition reference implementation",
"comparator": "flashinfer_bench_eval"
},
"measurement": {
"timer": "harness_reported",
"primaryStatistic": "mean"
},
"comparability": {
"family": "flashinfer_bench_trace"
}
},
"metadata": {
"name": "flashinfer-bench-trace",
"title": "FlashInfer-Bench trace harness"
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"environment": {
"kind": "ExecutionEnvironment",
"spec": {
"hardware": {
"vendor": "nvidia",
"product": "NVIDIA B200",
"architecture": "unknown"
},
"software": {
"framework": {
"name": "pytorch",
"version": "2.11.0+cu130"
},
"libraries": {
"triton": "3.6.0"
},
"cudaToolkit": "13.0"
}
},
"metadata": {
"name": "sol-trace-nvidia-b200-2-11-0-cu130-3-6-0-13-0",
"title": "NVIDIA B200 (trace-reported software)"
},
"apiVersion": "kernelindex.dev/v1alpha1"
}
}Cite this record (permalink, digest, access date)
Report an issue with this run
Published 2026-08-16 · FlashInfer-Bench · Apache-2.0All results for GEMM FP8 fp4 n4096 k4096 →JSON