Skip to content
Server

CUDA: Optimize reduce_rows_f32 kernel, leading up to 25x perf improvement on kernel-level and 10% perf increase for Gemma3n #17276

Sign in to view logs

CUDA: Optimize reduce_rows_f32 kernel, leading up to 25x perf improvement on kernel-level and 10% perf increase for Gemma3n

CUDA: Optimize reduce_rows_f32 kernel, leading up to 25x perf improvement on kernel-level and 10% perf increase for Gemma3n #17276

Triggered via pull request August 7, 2025 12:31
@ORipplerORippler
synchronize #15132
ORippler:osimons/optimize_reduce_rows_f32
Status Success
Total duration 11m 57s
Artifacts –

server.yml

on: pull_request
server-windows
7m 12s
server-windows
Matrix: server
Fit to window
Zoom out
Zoom in