torch.compile (inductor)
Conv2d Min Add Multiply · width = 128 · height = 128 · batch_size = 128 · in_channels = 64 · fp32 · run 01a03f0e-f84c…
●passedReported evidence
Not re-observed recently.
Primary measurement
3.08ms±0.00 · mean of 100
Rank 1 in its comparison group · source-native comparison · observed 2026-03-05
max ms 3.12 · min ms 3.07 · std ms 0.0117 · mean ms 3.08
Identity
implementationtorch.compile (inductor)
projectPyTorch
revisionunknown
workloadwidth = 128 · height = 128 · batch_size = 128 · in_channels = 64 · fp32
comparison keysha256:3bf85e32aaabd8ae…
sourceKernelBench baseline timings
external idH100_PCIe_LambdaLabs/torch-compile-inductor/level2/31_Conv2d_Min_Add_Multiply.py
sha256:7e8ec31c88db148f6d1ba1c…
Correctness
Marked passed by the source; the correctness policy was not published.
Workload
width128
height128
batch_size128
in_channels64
input_1fp32 [128, 64, 128, 128]
definition comparatornot_asserted
Measurements
latency · max3.12 ms · n=100
latency · mean3.08 ms · n=100
latency · min3.07 ms · n=100
latency · std11.7 µs
Protocol
harnessKernelBench timing scripts
timercuda_events
primaryStatisticmean
comparabilityFamilykernelbench_baseline_timing
Environment
gpuNVIDIA H100 (sm_90)
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": 3080000,
"maximum": 3120000,
"minimum": 3070000,
"confidence95": [
3077707,
3082293
]
},
"primaryStatistic": "mean"
},
"observedAt": "2026-03-05T08:38:15.000Z",
"measurements": [
{
"unit": "ns",
"value": 11700,
"metric": "latency",
"statistic": "std"
}
],
"sourceNative": {
"source": "kernelbench",
"metrics": {
"max_ms": 3.12,
"min_ms": 3.07,
"std_ms": 0.0117,
"mean_ms": 3.08
},
"benchmark": "H100_PCIe_LambdaLabs/baseline_time_torch_compile_inductor_default.json",
"externalId": "H100_PCIe_LambdaLabs/torch-compile-inductor/level2/31_Conv2d_Min_Add_Multiply.py"
},
"protocolDigest": "sha256:547725b033a9911eae0c5ac362344a85aed824894ed4ab476af945b9bb491ea6",
"workloadDigest": "sha256:31400b752d17dff6cdd9571c21650b45ff2fcdfdd681f943141318bca43eedf7",
"environmentDigest": "sha256:89f9b28b06758f22a1df331fea58bfde2c9cc0f3627d1829509743ea538f0011",
"implementationDigest": "sha256:625a45dcd6660d51ce1393c210b9204c3f4f8468a7578ade1e359c9f2ec2094d"
},
"metadata": {
"name": "kernelbench-h100-pcie-lambdalabs-torch-compile-inductor-l2-31-conv2d-min-add-multiply",
"title": "Conv2d Min Add Multiply · torch.compile (inductor) · Lambda Labs"
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"protocol": {
"kind": "BenchmarkProtocol",
"spec": {
"harness": {
"name": "KernelBench timing scripts",
"repository": "https://github.com/ScalingIntelligence/KernelBench"
},
"measurement": {
"timer": "cuda_events",
"primaryStatistic": "mean"
},
"comparability": {
"notes": "Mean of 100 timed forward passes (CUDA events, warm-up excluded) of the reference module on fixed inputs; the 95% interval is the mean's, from the reported standard deviation. The upstream JSON records no torch or CUDA version. Comparable only within one problem and one timing host.",
"family": "kernelbench_baseline_timing"
}
},
"metadata": {
"name": "kernelbench-timing-v1",
"title": "KernelBench baseline timing"
},
"apiVersion": "kernelindex.dev/v1alpha1"
},
"environment": {
"kind": "ExecutionEnvironment",
"spec": {
"hardware": {
"vendor": "nvidia",
"product": "NVIDIA H100",
"formFactor": "PCIe",
"architecture": "sm_90"
},
"software": {}
},
"metadata": {
"name": "kernelbench-h100-pcie-lambdalabs",
"title": "KernelBench host · H100 PCIe (Lambda Labs)"
},
"apiVersion": "kernelindex.dev/v1alpha1"
}
}Cite this record (permalink, digest, access date)
Report an issue with this run
Published 2026-08-26 · KernelBench baseline timings · MITAll results for Conv2d Min Add Multiply →JSON