优化ReduceSum和SortPair算子的吞吐量 #11

Closed
Ran128 wants to merge 5 commits from Ran128/GPUKernelContest:main into main
First-time contributor

本PR提供了两个关键GPU算子的高性能实现:

ReduceSum:通过两级归约策略和动态资源配置,在保证数值精度的同时,显著提升了归约计算的性能。

SortPair:采用三策略混合排序算法,根据数据规模自适应选择最优排序路径,实现了高效稳定的键值对排序。

赛题ID:6
参赛链接:https://gitlink.org.cn/Ran128/GPUKernelContest/tree/main/S1/6

本PR提供了两个关键GPU算子的高性能实现: ReduceSum:通过两级归约策略和动态资源配置,在保证数值精度的同时,显著提升了归约计算的性能。 SortPair:采用三策略混合排序算法,根据数据规模自适应选择最优排序路径,实现了高效稳定的键值对排序。 赛题ID:6 参赛链接:[https://gitlink.org.cn/Ran128/GPUKernelContest/tree/main/S1/6](https://gitlink.org.cn/Ran128/GPUKernelContest/tree/main/S1/6)
Ran128 added 4 commits 2025-12-07 17:23:48 +08:00
Ran128 added 1 commit 2025-12-07 17:27:19 +08:00
wawahejun closed this pull request 2026-02-13 17:58:10 +08:00

Pull request closed

Sign in to join this conversation.
No reviewers
No Label
No Milestone
No project
No Assignees
1 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: ccf-ai-infra/GPUKernelContest#11
No description provided.