Skip to content

Fix #1638: Faster returning of results - #1647

Merged
mdboom merged 2 commits into
NVIDIA:mainfrom
mdboom:fast-result
Feb 18, 2026
Merged

Fix #1638: Faster returning of results#1647
mdboom merged 2 commits into
NVIDIA:mainfrom
mdboom:fast-result

Conversation

@mdboom

@mdboommdboom commented Feb 18, 2026

Copy link
Copy Markdown
Contributor

See #1638 for details as to why this works.

On my machine with the #659 benchmark, I see a reduction from 3.02us to 2.54us per iteration.

@copy-pr-bot

copy-pr-botBot commented Feb 18, 2026

Copy link
Copy Markdown
Contributor

Auto-sync is disabled for ready for review pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

@mdboommdboom self-assigned this Feb 18, 2026
@mdboom

mdboom commented Feb 18, 2026

Copy link
Copy Markdown
ContributorAuthor

/ok to test

@github-actions

This comment has been minimized.

Sign in to view
cpcloud
cpcloud approved these changes Feb 18, 2026
View reviewed changes

@cpcloudcpcloud left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems legit!

@mdboommdboom added cuda.bindings Everything related to the cuda.bindings module performance labels Feb 18, 2026
@mdboom
mdboom merged commit 7a73c60 into NVIDIA:mainFeb 18, 2026
93 checks passed
@github-actions

github-actionsBot commented Feb 18, 2026

Copy link
Copy Markdown
Doc Preview CI
Preview removed because the pull request was closed or merged.

@leofang

leofang commented Feb 18, 2026

Copy link
Copy Markdown
Member

@mdboom sorry for noticing this late, but I am confused by the diff -- it seems like this PR just fixes a recently-introduced (I think) regression? We've fixed this enum issue long time ago by a fast dict lookup (#546), but that fix seems to be gone (and replaced by a presumably even-faster memoization in this PR).

@mdboommdboom mentioned this pull request Feb 19, 2026
Merged
@leofang

leofang commented Feb 19, 2026•
edited
Loading

Copy link
Copy Markdown
Member

it seems like this PR just fixes a recently-introduced (I think) regression?

Ah, I see. The "regression" was introduced in the fast enum refactoring (#1581). So with this PR it means regardless of how fast the enum implementation (builtin or custom) is the memoization is always needed... 🙂

@rwgkrwgk mentioned this pull request Feb 27, 2026
Merged
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

1 more reviewer
@cpcloudcpcloudcpcloud approved these changes
Reviewers whose approvals may not affect merge requirements

Assignees

@mdboommdboom

Labels

cuda.bindingsEverything related to the cuda.bindings moduleperformance

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@mdboom@leofang@cpcloud@rparolin