Search · KernelIndex

Fastest reported

20.2µsgeomean

submission 319574_spatters · License unknown · python

Reported evidence · last observed 2026-01-10. Reported by source; not independently reproduced. License unknown.

No package · vendor the source →View source →Run detail →

Ranked under the source's published protocolWhy comparable?

GPUNVIDIA B200

Workloadsuite of 4 cases · geomean score

Protocolreference-kernels eval.py · geomean

The query left these dimensions open; this is the most-measured of 2 cohorts. The hardware chips above switch it.

Exact 2 Other protocols 1

Has source1License known0Installable0

Measured under a different protocol. Shown as published, never ranked against exact rows.

Implementation

Latency

Trust

Observed

—

submission 370774_spatters

20.8µs

Reported · license unknown · source

2026-01-18stale

Reported by source; not independently reproduced. Different comparison cohort: protocol or environment differs.

source mirroredlicense unknownno install recipeView source →Run detail →Compare with #1 →

3D tensor matrix multiplicationOperation in the gemm family4D tensor matrix multiplicationOperation in the gemm familyMatmul with diagonal matricesOperation in the gemm familyMatmul for symmetric matricesOperation in the gemm familyMatmul for upper triangular matricesOperation in the gemm familyMatmul for lower triangular matricesOperation in the gemm family

Missing a kernel here? Evidence submissions open with the contribution beta.

This search as an API call

curlkicomparison key sha256:25ad78ee1791ba39…JSON

GPU MODE KernelBot · last observed 2026-01-20