submission 812268
Michael Ha · python · License unknown
Use it
Vendorable · source mirrored · license unknownView source →
No package. Vendor the mirrored source: 13 lines, June 9 Researcher Reciprocity License v1.0.
test.py
curl "https://kernelindex.com/api/v1/implementations/kernelbot-qr-v2-812268?include=source"interfacepython
Compatibility
measured onNVIDIA B200
declared hardwareNVIDIA B200
architecturessm_100
dtypesfp32
Benchmark evidence
1 measurement across 1 GPU, fastest first.
Operation / workload
Hardware
Latency
Rank
Observed
Reported · How evidence levels are derived →
Source and license
sourceavailable
revision digestsha256:b2bea2bd597cf8330430b885ad7de65b63aed715a7b8ce1f6edeeac3b55ee8ee
license declaredunknown
license concludedunknown
authorsMichael Ha
imported2026-08-26
Kernel source
test.py13 lines
import torch
from task import input_t, output_t
def custom_kernel(data: input_t) -> output_t:
"""
Fast batched square compact-Householder QR factorization.
Delegates to PyTorch's highly optimized cuSOLVER-backed torch.geqrf.
This ensures strict FP32 correctness against the LAPACK-style checker
while leveraging hardware-accelerated batched GEMM/QR kernels.
"""
return torch.geqrf(data)Source code from GPU Mode and the KernelBot dataset · June 9 Researcher Reciprocity License v1.0
Best evidence level for this revision: reported
JSON