submission 673601
Elán Zainos Corona · python · License unknown
Use it
Vendorable · source mirrored · license unknownView source →
No package. Vendor the mirrored source: 17 lines, June 9 Researcher Reciprocity License v1.0.
VectorSumR1.py
curl "https://kernelindex.com/api/v1/implementations/kernelbot-vectorsum-v2-673601?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 digestsha256:7b996f58a86276e32c0c3a2be39e65c26c413becd24ed90db6a64eabe2eef080
license declaredunknown
license concludedunknown
authorsElán Zainos Corona
imported2026-08-15
Kernel source
VectorSumR1.py17 lines
# -*- coding: utf-8 -*-
import os
import torch
# Estabilización absoluta del determinismo
os.environ["CUBLAS_WORKSPACE_CONFIG"] = ":4096:8"
torch.use_deterministic_algorithms(True)
def custom_kernel(data):
"""
Sentinel Omega - Motor de Reducción de Vértices.
Implementación de ultra-baja latencia para 43µs.
"""
# Unpacking de alto flujo (data[1] se ignora para colapso escalar)
# Usar .sum() con el parámetro dtype es el "estándar de oro".
return data[0].sum(dtype=torch.float64).to(torch.float32)
Source code from GPU Mode and the KernelBot dataset · June 9 Researcher Reciprocity License v1.0
Best evidence level for this revision: reported
JSON