submission 19663
AMD FP8 blockwise GEMM · k = 256 · m = 1024 · n = 7168 · seed = 5364 · fp32/fp8_e4m3fnuz · run 01a006cd-66c3…
●passedReported evidence
Not re-observed recently.
Primary measurement
78.0µsmean of 100
Rank 85 in its comparison group · source-native comparison · observed 2025-05-10
best ns 76082 · mean ns 78026.87 · leaderboard score s 0.0004548239232360519
Identity
implementationsubmission 19663
projectKernelBot user 334
revisionsubmission-19663
workloadk = 256 · m = 1024 · n = 7168 · seed = 5364 · fp32/fp8_e4m3fnuz
comparison keysha256:66b4440706ab6298…
sourceGPU MODE KernelBot
external idsubmission/19663/benchmark/3
sha256:2b3cdb09de305481346add8…
Correctness
Marked passed by the source; the correctness policy was not published.
Workload
k256
m1024
n7168
seed5364
afp8_e4m3fnuz [1024, 256]
bfp8_e4m3fnuz [7168, 256]
a_scalefp32 [1024, 2]
b_scalefp32 [56, 2]
definition comparatorkernelbot_eval
Measurements
latency · max84.4 µs · n=100
latency · mean78.0 µs · n=100
latency · min76.1 µs · n=100
latency · std1.26 µs · n=100
Protocol
harnessreference-kernels eval.py
timerwall_clock
primaryStatisticmean
correctnessReferenceproblem reference implementation
comparabilityFamilygpumode_kernelbot
Environment
gpuAMD Instinct MI300X (gfx942)
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": {
"samples": 100,
"latencyNs": {
"mean": 78027,
"maximum": 84355,
"minimum": 76082
},
"primaryStatistic": "mean"
},
"observedAt": "2025-05-10T15:20:16.878Z",
"measurements": [
{
"unit": "ns",
"value": 1257.3943287930042,
"metric": "latency",
"statistic": "std",
"sampleCount": 100
}
],
"sourceNative": {
"source": "gpumode-kernelbot",
"metrics": {
"best_ns": 76082,
"mean_ns": 78026.87,
"leaderboard_score_s": 0.0004548239232360519
},
"benchmark": "amd-fp8-mm/3",
"externalId": "submission/19663"
},
"protocolDigest": "sha256:7c66da4b5866502eb10444a7ba013d0f6bdf1620c55847945008cad2fbf98aec",
"workloadDigest": "sha256:25d9eb98d2e49165b9b36ae40dffda6a963353e9b6dc3308cbb05a76da45af7d",
"environmentDigest": "sha256:f722892449ed976f3e21b12df8b92b34573fc232883fec6a2f3da399a00abdd5",
"implementationDigest": "sha256:6379da484b9c90a1bab219907648d8127ed9a6b64ad240d8c5da9ab7391b3363"
},
"metadata": {
"name": "kernelbot-19663-k256-m1024-n7168-seed5364",
"title": "amd-fp8-mm · submission 19663 · case 3",
"labels": {
"torch": "2.8.0.dev20250509+rocm6.3",
"platform": "Linux-5.15.0-136-generic-x86_64-with-glibc2.35"
}
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"protocol": {
"kind": "BenchmarkProtocol",
"spec": {
"harness": {
"name": "reference-kernels eval.py",
"repository": "https://github.com/gpu-mode/reference-kernels"
},
"correctness": {
"reference": "problem reference implementation",
"comparator": "kernelbot_eval"
},
"measurement": {
"timer": "wall_clock",
"primaryStatistic": "mean"
},
"comparability": {
"notes": "Per-case wall-clock statistics as published in the KernelBot dataset. Comparable only within one leaderboard problem and GPU type; runner software versions ride each run's labels.",
"family": "gpumode_kernelbot"
}
},
"metadata": {
"name": "gpumode-kernelbot-eval",
"title": "GPU MODE KernelBot evaluation harness"
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"environment": {
"kind": "ExecutionEnvironment",
"spec": {
"hardware": {
"vendor": "amd",
"product": "AMD Instinct MI300X",
"architecture": "gfx942"
},
"software": {}
},
"metadata": {
"name": "gpumode-kernelbot-amd-instinct-mi300x",
"title": "KernelBot runner · AMD Instinct MI300X"
},
"apiVersion": "kernelindex.dev/v1alpha1"
}
}Cite this record (permalink, digest, access date)
Report an issue with this run
Published 2026-08-15 · GPU MODE KernelBot · June 9 Researcher Reciprocity License v1.0All results for AMD FP8 blockwise GEMM →JSON