New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Improve n-dimensional unique performance #8328

Draft

andfoy wants to merge 3 commits into cupy:main from andfoy:improve_unique_axis

+492 −91

Contributor

andfoy commented May 17, 2024

Fixes #8326

This PR aims to improve the performance of cupy.unique when axis != None, the main idea goes as follows:

Compute a hash value for each row. This can be done completely in parallel
Call unique_1d, which should yield the unique rows according to their hash value.


          Start working on improvements to unique over axis

a4a5966

takagi self-assigned this

takagi added cat:enhancement prio:medium labels

Member

leofang commented May 20, 2024

cc: @essoca for vis

andfoy added 2 commits

May 23, 2024 19:36


          Reduce CRC alongside axis=1

a52712a


          Extend CRC to other dtypes

f4c3a69

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment