Run 01a02032 · KernelIndex

Primary measurement

78.1µsmean

Rank 3 in its comparison group · source-native comparison · observed 2025-10-20

Compare with #1 →

66.31× reference

Identity

implementationgpt-o3_cuda_9abd34

projectgpt-o3

revisionda915083d4c7c5e61aa3005e3d17ae488e0fc71c

workloadtotal_q = 6 · total_kv = 6 · len_indptr = 2 · bf16

comparison keysha256:1e601cad8d79776a…

sourceFlashInfer-Bench

external idtrace/gpt-o3_cuda_9abd34/ebf7188b-4b31-4746-b57b-fa25b53f5e3e

sha256:4b5ebf8d59aae0bab7439ee…

Correctness

comparatorflashinfer_bench_eval

max absolute error0.0000152587890625

max relative error0.005952357780188322

resultpassed

Workload

total_q6

total_kv6

len_indptr2

kbf16 [6, 4, 128]

qbf16 [6, 32, 128]

vbf16 [6, 4, 128]

kv_indptrint32 [2]

qo_indptrint32 [2]

definition comparatorsol_execbench_eval

Measurements

latency · mean78.1 µs

Protocol

harnessflashinfer-bench

timerharness_reported

primaryStatisticmean

correctnessReferencedefinition reference implementation

comparabilityFamilyflashinfer_bench_trace

Environment

gpuNVIDIA B200 (unknown)

cudaToolkit12.8

frameworkpytorch 2.8.0+cu128

Artifacts

logssha256:404326344684e4de… · text/plain · unavailable

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": 78059
        },
        "primaryStatistic": "mean"
      },
      "evidence": {
        "logs": {
          "uri": "kernelindex:artifact/sha256:404326344684e4de581e4e71f1f2412c92ca6916fd830b2aca778682571fe80f",
          "digest": "sha256:404326344684e4de581e4e71f1f2412c92ca6916fd830b2aca778682571fe80f"
        }
      },
      "observedAt": "2025-10-20T01:06:27.212Z",
      "correctness": {
        "comparator": "sol_execbench_eval",
        "maximumAbsoluteError": 0.0000152587890625,
        "maximumRelativeError": 0.005952357780188322
      },
      "sourceNative": {
        "source": "flashinfer-bench",
        "metrics": {
          "latency_ms": 0.07805945052275558,
          "speedup_factor": 66.30853729312169,
          "reference_latency_ms": 5.176007986068726
        },
        "benchmark": "gqa_ragged_prefill_causal_h32_kv4_d128"
      },
      "protocolDigest": "sha256:7d44d206ad086b4bece35ce31f6531ebb11f07e0d9dcfb74c75be86d3ec02095",
      "workloadDigest": "sha256:6d66441876c07069445f683eb915c8686bd27a086dae10c521fbd12247c46d3f",
      "environmentDigest": "sha256:f4292a8be127b7230cd3995c1d605fb6c644e2c2f94566e1d28cc1be766ca5d5",
      "implementationDigest": "sha256:3a73a83021dd9abc23c3f9cf84a55ebd359eadfde4346501fc90586dd7f3116c"
    },
    "metadata": {
      "name": "sol-trace-gqa-ragged-prefill-causal-h32-kv4-d128-ebf7188b",
      "title": "gpt-o3_cuda_9abd34 · gqa_ragged_prefill_causal_h32_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": "2.8.0+cu128"
        },
        "libraries": {
          "triton": "3.4.0"
        },
        "cudaToolkit": "12.8"
      }
    },
    "metadata": {
      "name": "sol-trace-nvidia-b200-2-8-0-cu128-3-4-0-12-8",
      "title": "NVIDIA B200 (trace-reported software)"
    },
    "apiVersion": "kernelindex.dev/v1alpha1"
  }
}

Cite this record (permalink, digest, access date)

Report an issue with this run

Published 2026-08-20 · FlashInfer-Bench · Apache-2.0All results for GQA ragged prefill causal h32 kv4 d128 →JSON

gpt-o3 / cuda9abd34