Skip to content

Conversation

@trivedivivek
Copy link
Contributor

@trivedivivek trivedivivek commented Oct 23, 2024

…, to reduce shader register pressure.

This diff reduces the precision of texture coordinates in the conv2d_pw op in Executorch's Vulkan backend to reduce shader register pressure. The changes made in the code include reducing the precision of the z coordinate in the loop and using uint16_t instead of int for the loop counter.

Differential Revision: [D64767415](https://our.internmc.facebook.com/intern/diff/D64767415/)

[ghstack-poisoned]
@pytorch-bot
Copy link

pytorch-bot bot commented Oct 23, 2024

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6468

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

✅ No Failures

As of commit 518a65f with merge base fa30e80 (image):
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D64767415

trivedivivek added a commit that referenced this pull request Oct 23, 2024
…, to reduce shader register pressure.

This diff reduces the precision of texture coordinates in the conv2d_pw op in Executorch's Vulkan backend to reduce shader register pressure. The changes made in the code include reducing the precision of the z coordinate in the loop and using uint16_t instead of int for the loop counter.

Differential Revision: [D64767415](https://our.internmc.facebook.com/intern/diff/D64767415/)

ghstack-source-id: 249679438
Pull Request resolved: #6468
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 Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants