Transformers
GroupNorm · batch = 128 · hidden = 512 · channels = 32 · channels_per_group = 4 · fp32 · run 01a0202b-234e…
●passedReported evidence
Not re-observed recently.
Primary measurement
84.0µsmedian
Rank 1 in its comparison group · source-native comparison · observed 2026-02-28
ms 20 0.08390399813652039 · ms 50 0.08396799862384796 · ms 80 0.08403199911117554
Identity
implementationTransformers
projectHugging Face Transformers
revisionunknown
workloadbatch = 128 · hidden = 512 · channels = 32 · channels_per_group = 4 · fp32
comparison keysha256:0d6322b132ec5373…
sourceLiger-Kernel benchmarks
external idgroup_norm/full/huggingface/nvidia-b200/batch128-channels32-channels_per_group4-hidden512-fp32/2026-02-28-00-23-26
sha256:fd112b1be07acdee93f0b49…
Correctness
Marked passed by the source; the correctness policy was not published.
Workload
batch128
hidden512
channels32
channels_per_group4
xfp32 [128, 32, 512]
biasfp32 [32]
weightfp32 [32]
definition comparatornot_asserted
Measurements
latency · median84.0 µs
latency · p2083.9 µs
latency · p8084.0 µs
Protocol
harnessLiger-Kernel benchmark scripts
timercuda_events
primaryStatisticmedian
comparabilityFamilyliger_kernel_bench
Environment
gpuNVIDIA B200 (sm_100)
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": {
"median": 83968
},
"primaryStatistic": "median"
},
"observedAt": "2026-02-28T00:23:26.000Z",
"measurements": [
{
"unit": "ns",
"value": 83904,
"metric": "latency",
"statistic": "p20"
},
{
"unit": "ns",
"value": 84032,
"metric": "latency",
"statistic": "p80"
}
],
"sourceNative": {
"source": "liger-kernel-bench",
"metrics": {
"ms_20": 0.08390399813652039,
"ms_50": 0.08396799862384796,
"ms_80": 0.08403199911117554
},
"benchmark": "group_norm/full",
"externalId": "group_norm/full/huggingface/nvidia-b200/batch128-channels32-channels_per_group4-hidden512-fp32/2026-02-28-00-23-26"
},
"protocolDigest": "sha256:5b02ee9a2140fbc6fcc1c89800f6c177ba22718c8897cb532bf0ba91867c1e58",
"workloadDigest": "sha256:7ca7a18b215eaf6a299e0a2b8ab2c3f1fc59e496869572f778794ab9e1664354",
"environmentDigest": "sha256:ad459e643ef802f7008f1fa8915785c905d05abbfdb5eb0328be214f8d9f3a8b",
"implementationDigest": "sha256:e8c7d555b6002524eb04f3e8a91c03603172925143b51877f4050be074e9e025"
},
"metadata": {
"name": "liger-group-norm-full-huggingface-nvidia-b200-batch128-channels32-channels-per-group4-hidden512-fp32",
"title": "GroupNorm · huggingface · full",
"labels": {
"liger_version": "0.7.0"
}
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"protocol": {
"kind": "BenchmarkProtocol",
"spec": {
"harness": {
"name": "Liger-Kernel benchmark scripts",
"repository": "https://github.com/linkedin/Liger-Kernel"
},
"measurement": {
"timer": "cuda_events",
"primaryStatistic": "median"
},
"comparability": {
"notes": "Timed: the forward pass plus the backward pass, via triton.testing.do_bench with quantiles 0.5/0.2/0.8 (median with p20/p80 spread). The upstream CSV records no CUDA, driver, or torch version; the Liger release rides each run's labels. Comparable only within one kernel, workload, timed pass, and GPU.",
"family": "liger_kernel_bench"
}
},
"metadata": {
"name": "liger-bench-do-bench-full",
"title": "Liger benchmark harness · full pass"
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"environment": {
"kind": "ExecutionEnvironment",
"spec": {
"hardware": {
"vendor": "nvidia",
"product": "NVIDIA B200",
"architecture": "sm_100"
},
"software": {}
},
"metadata": {
"name": "liger-bench-nvidia-b200",
"title": "Liger benchmark host · B200"
},
"apiVersion": "kernelindex.dev/v1alpha1"
}
}Cite this record (permalink, digest, access date)
Report an issue with this run
Published 2026-08-20 · Liger-Kernel benchmarks · BSD-2-ClauseAll results for GroupNorm →JSON