Skip to content
KernelIndex
Search⌘K

submission 511695

David Holtz · python · License unknown

Use it

Vendorable · source mirrored · license unknownView source →

No package. Vendor the mirrored source: 16 lines, June 9 Researcher Reciprocity License v1.0.

vectorsum.py
curl "https://kernelindex.com/api/v1/implementations/kernelbot-vectorsum-v2-511695?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
Vector sum reductionsuite of 6 cases
NVIDIA B200
47.0µs
#17 of 88
2026-02-27

Reported · How evidence levels are derived →

Source and license

sourceavailable
revision digestsha256:dc1adbbe4927b53934534761ab989b3a167045d6c83f7f21e0fc07cd958cde01
license declaredunknown
license concludedunknown
authorsDavid Holtz
imported2026-08-15

Kernel source

vectorsum.py16 lines
import subprocess
import sys

for pkg in ["kernels"]:
    subprocess.check_call([sys.executable, "-m", "pip", "install", "--upgrade", pkg])

from kernels import get_kernel

kernel = get_kernel("drbh/vectorsum-dev", revision="e7f67b2e9b6ddc5b157f02cc90dc3a8276b83dd8")
_vecsum = kernel.VectorSum("cuda")


def custom_kernel(data):
    x, _ = data
    return _vecsum(x)  # already scalar, no squeeze

Source code from GPU Mode and the KernelBot dataset · June 9 Researcher Reciprocity License v1.0

Changes from previous submission

Against this author's previous submission submission 511683.

Best evidence level for this revision: reported

JSON