flashinfer_wrapper_023122
GQA ragged prefill causal h20 kv4 d128 · total_q = 1600 · total_kv = 1600 · len_indptr = 33 · bf16 · run 01a00b1e-2504…
●passedReported evidence
Not re-observed recently.
Primary measurement
95.9µsmean
Rank 1 in its comparison group · source-native comparison · observed 2026-03-24
128.09× reference
Identity
implementationflashinfer_wrapper_023122
projectFlashInfer-Bench baselines
revisionda915083d4c7c5e61aa3005e3d17ae488e0fc71c
workloadtotal_q = 1600 · total_kv = 1600 · len_indptr = 33 · bf16
comparison keysha256:84473fa0f5c3a922…
sourceFlashInfer-Bench
external idtrace/flashinfer_wrapper_023122/f26e853b-dd4f-4e13-86cf-51f9fdc54e0f
sha256:54737a45a82da382af21765…
Correctness
comparatorflashinfer_bench_eval
max absolute error0.03125
max relative error1944.255615234375
resultpassed
Workload
total_q1600
total_kv1600
len_indptr33
kbf16 [1600, 4, 128]
qbf16 [1600, 20, 128]
vbf16 [1600, 4, 128]
kv_indptrint32 [33]
qo_indptrint32 [33]
definition comparatorsol_execbench_eval
Measurements
latency · mean95.9 µs
Protocol
harnessflashinfer-bench
timerharness_reported
primaryStatisticmean
correctnessReferencedefinition reference implementation
comparabilityFamilyflashinfer_bench_trace
Environment
gpuNVIDIA B200 (unknown)
cudaToolkit12.8
frameworkpytorch 2.9.1+cu128
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": 95937
},
"primaryStatistic": "mean"
},
"observedAt": "2026-03-24T20:24:21.500Z",
"correctness": {
"comparator": "sol_execbench_eval",
"maximumAbsoluteError": 0.03125,
"maximumRelativeError": 1944.255615234375
},
"sourceNative": {
"source": "flashinfer-bench",
"metrics": {
"latency_ms": 0.095937,
"speedup_factor": 128.08744280100484,
"reference_latency_ms": 12.288325
},
"benchmark": "gqa_ragged_prefill_causal_h20_kv4_d128"
},
"protocolDigest": "sha256:7d44d206ad086b4bece35ce31f6531ebb11f07e0d9dcfb74c75be86d3ec02095",
"workloadDigest": "sha256:afb70137548b428d32aa44d60201da541083ed01b00270df5d74a795064ccf8c",
"environmentDigest": "sha256:af7cb68a0bb0eb9671092a620668c2600f0431301fc46a71807ea75fb7de8bce",
"implementationDigest": "sha256:9fdf07cc4db985f53d8d56be273c9c8277b2168d3e9337f5eea32a574764b251"
},
"metadata": {
"name": "sol-trace-gqa-ragged-prefill-causal-h20-kv4-d128-f26e853b",
"title": "flashinfer_wrapper_023122 · gqa_ragged_prefill_causal_h20_kv4_d128"
},
"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": "2.9.1+cu128"
},
"libraries": {
"triton": "3.5.1"
},
"cudaToolkit": "12.8"
}
},
"metadata": {
"name": "sol-trace-nvidia-b200-2-9-1-cu128-3-5-1-12-8",
"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 GQA ragged prefill causal h20 kv4 d128 →JSON