Run 01a00b1d · KernelIndex

Primary measurement

179.8µsmean

Rank 1 in its comparison group · source-native comparison · observed 2026-04-12

10.75× reference

Identity

implementationflashinfer_wrapper_b38b5f

projectFlashInfer-Bench baselines

revisionda915083d4c7c5e61aa3005e3d17ae488e0fc71c

workloadtotal_q = 622 · num_pages = 5557 · len_indptr = 4 · num_kv_indices = 622 · bf16

comparison keysha256:3ba958e2f89ab534…

sourceFlashInfer-Bench

external idtrace/flashinfer_wrapper_b38b5f/ac8b7b21-09d1-46e0-b724-e70d3ccee8cb

sha256:f580be2ab0b67067403a2cc…

Correctness

comparatorflashinfer_bench_eval

max absolute error0.015625

max relative error3100.836669921875

resultpassed

Workload

total_q622

num_pages5557

len_indptr4

num_kv_indices622

qbf16 [622, 24, 128]

k_cachebf16 [5557, 1, 8, 128]

v_cachebf16 [5557, 1, 8, 128]

kv_indptrint32 [4]

qo_indptrint32 [4]

kv_indicesint32 [622]

definition comparatorsol_execbench_eval

Measurements

latency · mean179.8 µs

Protocol

harnessflashinfer-bench

timerharness_reported

primaryStatisticmean

correctnessReferencedefinition reference implementation

comparabilityFamilyflashinfer_bench_trace

Environment

gpuNVIDIA GB200 (unknown)

cudaToolkit13.0

frameworkpytorch 2.9.1+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 manifest

{
  "run": {
    "kind": "BenchmarkRun",
    "spec": {
      "status": "passed",
      "timing": {
        "latencyNs": {
          "mean": 179792
        },
        "primaryStatistic": "mean"
      },
      "observedAt": "2026-04-12T12:30:39.975Z",
      "correctness": {
        "comparator": "sol_execbench_eval",
        "maximumAbsoluteError": 0.015625,
        "maximumRelativeError": 3100.836669921875
      },
      "sourceNative": {
        "source": "flashinfer-bench",
        "metrics": {
          "latency_ms": 0.1797919993599256,
          "speedup_factor": 10.75043760986176,
          "reference_latency_ms": 1.9328426718711853
        },
        "benchmark": "gqa_paged_prefill_causal_h24_kv8_d128_ps1"
      },
      "protocolDigest": "sha256:7d44d206ad086b4bece35ce31f6531ebb11f07e0d9dcfb74c75be86d3ec02095",
      "workloadDigest": "sha256:51f770da86b83438c32828d5b31568853c18d81ddc0487c03605d102085eca1f",
      "environmentDigest": "sha256:1766500d1b1ecc9787a211f3c98e76bf311671f443544a305252cb4d9564be8b",
      "implementationDigest": "sha256:29979918bd047f743af3d32549b27f45f75afc469627e5a41df4aee474db5f90"
    },
    "metadata": {
      "name": "sol-trace-gqa-paged-prefill-causal-h24-kv8-d128-ps1-ac8b7b21",
      "title": "flashinfer_wrapper_b38b5f · gqa_paged_prefill_causal_h24_kv8_d128_ps1"
    },
    "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 GB200",
        "architecture": "unknown"
      },
      "software": {
        "framework": {
          "name": "pytorch",
          "version": "2.9.1+cu130"
        },
        "libraries": {
          "triton": "3.3.0"
        },
        "cudaToolkit": "13.0"
      }
    },
    "metadata": {
      "name": "sol-trace-nvidia-gb200-2-9-1-cu130-3-3-0-13-0",
      "title": "NVIDIA GB200 (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 prefill causal h24 kv8 d128 ps1 →JSON

flashinfer / wrapperb38b5f