Skip to content
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.

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

Use alignmend memory for DrJit error functions #98

Open
wants to merge 2 commits into
base: main
Choose a base branch
from

Conversation

jeongseok-meta
Copy link
Contributor

Summary: This diff updates the DrJit implementation of the SIMD error functions to utilize the aligned memory versions of DrJit by correctly offsetting the Jacobian pointer to the aligned memory, similar to the AVX implementations.

Differential Revision: D64133416

Summary:
Memory alignment checks were added by D63658278 for both AVX and DrJit implementations, but introduced an error in the AVX implementation where Jacobian pointers were not offset to aligned memory address.

Diff fixes issue by using new utility function that computes offset and checks alignment with offset, and also removes alignment check from DrJit implementation since current implementation does not require memory alignment (to be addressed in upper Diff stack).

Differential Revision: D64132292
Summary: This diff updates the DrJit implementation of the SIMD error functions to utilize the aligned memory versions of DrJit by correctly offsetting the Jacobian pointer to the aligned memory, similar to the AVX implementations.

Differential Revision: D64133416
@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Meta Open Source bot. label Oct 9, 2024
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D64133416

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Meta Open Source bot. fb-exported
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants