submission 512318

Use it

Vendorable · source mirrored · license unknownView source →

Benchmark evidence

1 measurement across 1 GPU, fastest first.

Reported · How evidence levels are derived →

Source and license

Kernel source

#!POPCORN leaderboard prefixsum_v2
#!POPCORN gpu A100

import torch
from task import input_t, output_t


def custom_kernel(data: input_t) -> output_t:
    x, _y = data
    return torch.cumsum(x, dim=0)

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 512313.

  ⋯ 5 unchanged lines
   
   
  def custom_kernel(data: input_t) -> output_t:
-     x, y = data
-     torch.cumsum(x, dim=0, out=y)
-     return y
+     x, _y = data
+     return torch.cumsum(x, dim=0)