flashinfer_wrapper_bcbabf
GQA paged decode h20 kv4 d128 ps64 · num_pages = 15292 · batch_size = 16 · len_indptr = 17 · num_kv_indices = 13578 · bf16 · run 01a00b1d-2e0b…
●passedReported evidence
Not re-observed recently.
Primary measurement
21.8msmean
Rank 2 in its comparison group · source-native comparison · observed 2026-03-28
24.51× reference
Identity
implementationflashinfer_wrapper_bcbabf
projectFlashInfer-Bench baselines
revisionda915083d4c7c5e61aa3005e3d17ae488e0fc71c
workloadnum_pages = 15292 · batch_size = 16 · len_indptr = 17 · num_kv_indices = 13578 · bf16
comparison keysha256:b485d6e496cdb59a…
sourceFlashInfer-Bench
external idtrace/flashinfer_wrapper_bcbabf/dbfd0b92-450f-425f-987a-d2ab665fdc3f
sha256:d53bd2b16e096dd4783a4bb…
Correctness
comparatorflashinfer_bench_eval
max absolute error0.0001220703125
max relative error48.2713623046875
resultpassed
Workload
num_pages15292
batch_size16
len_indptr17
num_kv_indices13578
qbf16 [16, 20, 128]
k_cachebf16 [15292, 64, 4, 128]
v_cachebf16 [15292, 64, 4, 128]
kv_indptrint32 [17]
kv_indicesint32 [13578]
kv_last_page_lenint32 [16]
definition comparatorsol_execbench_eval
Measurements
latency · mean21.8 ms
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": 21795042
},
"primaryStatistic": "mean"
},
"observedAt": "2026-03-28T04:31:28.996Z",
"correctness": {
"comparator": "sol_execbench_eval",
"maximumAbsoluteError": 0.0001220703125,
"maximumRelativeError": 48.2713623046875
},
"sourceNative": {
"source": "flashinfer-bench",
"metrics": {
"latency_ms": 21.795042,
"speedup_factor": 24.507523224777454,
"reference_latency_ms": 534.142498
},
"benchmark": "gqa_paged_decode_h20_kv4_d128_ps64"
},
"protocolDigest": "sha256:7d44d206ad086b4bece35ce31f6531ebb11f07e0d9dcfb74c75be86d3ec02095",
"workloadDigest": "sha256:80f440957b2acec749aedfb26a7614ec841c3973b0ae516e139a8ab749f90337",
"environmentDigest": "sha256:af7cb68a0bb0eb9671092a620668c2600f0431301fc46a71807ea75fb7de8bce",
"implementationDigest": "sha256:1a081dab2e7f927149c8b3aa27e3bd7d401852544e3b7e2d2d8f6dddd0901d1c"
},
"metadata": {
"name": "sol-trace-gqa-paged-decode-h20-kv4-d128-ps64-dbfd0b92",
"title": "flashinfer_wrapper_bcbabf · gqa_paged_decode_h20_kv4_d128_ps64"
},
"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 paged decode h20 kv4 d128 ps64 →JSON