Liger
Jensen-Shannon divergence loss · vocab = 128256 · tokens = 2048 · fp32 · run 01a016af-76cc…
●passedReported evidence
Primary measurement
9.11msmedian
Rank 2 in its comparison group · source-native comparison · observed 2026-05-27
ms 20 9.110560417175293 · ms 50 9.110560417175293 · ms 80 9.110560417175293
Identity
implementationLiger
projectLiger-Kernel
revisionunknown
workloadvocab = 128256 · tokens = 2048 · fp32
comparison keysha256:26a3d326c83203eb…
sourceLiger-Kernel benchmarks
external idjsd/forward/liger/nvidia-b200/tokens2048-vocab128256-fp32/2026-05-27-17-14-29
sha256:efaaa145f23845f43342a26…
Correctness
Marked passed by the source; the correctness policy was not published.
Workload
vocab128256
tokens2048
student_log_probsfp32 [2048, 128256]
teacher_log_probsfp32 [2048, 128256]
definition comparatornot_asserted
Measurements
latency · median9.11 ms
latency · p209.11 ms
latency · p809.11 ms
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": 9110560
},
"primaryStatistic": "median"
},
"observedAt": "2026-05-27T17:14:29.000Z",
"measurements": [
{
"unit": "ns",
"value": 9110560,
"metric": "latency",
"statistic": "p20"
},
{
"unit": "ns",
"value": 9110560,
"metric": "latency",
"statistic": "p80"
}
],
"sourceNative": {
"source": "liger-kernel-bench",
"metrics": {
"ms_20": 9.110560417175293,
"ms_50": 9.110560417175293,
"ms_80": 9.110560417175293
},
"benchmark": "jsd/forward",
"externalId": "jsd/forward/liger/nvidia-b200/tokens2048-vocab128256-fp32/2026-05-27-17-14-29"
},
"protocolDigest": "sha256:a79eaf4df66c97c6adc1fb2d09df78682155d1ea86652ae0b70199875bde7747",
"workloadDigest": "sha256:744bcd0ae3327229f524f2b47b751b38fad004a729355e2411ce31caf32f71d9",
"environmentDigest": "sha256:ad459e643ef802f7008f1fa8915785c905d05abbfdb5eb0328be214f8d9f3a8b",
"implementationDigest": "sha256:cfb3edb701b526d49e58d1de743fa6b7ce8e9b2d82736c8a68b45577428f96fc"
},
"metadata": {
"name": "liger-jsd-forward-liger-nvidia-b200-tokens2048-vocab128256-fp32-2026-05-27-17-14-29",
"title": "Jensen-Shannon divergence loss · liger · forward",
"labels": {
"liger_version": "0.8.0"
}
},
"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 module forward 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-forward",
"title": "Liger benchmark harness · forward 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 2026-08-18 · Liger-Kernel benchmarks · BSD-2-ClauseAll results for Jensen-Shannon divergence loss →JSON