[ET-VK] Reduced int precision for all int storage in q linear op, and reducing some texture coordinate storage variables to improve performance.

[trivedivivek]

Stack from ghstack (oldest at bottom):

• [ET-VK] Reduced int precision for texture coordinates in conv2d_pw op, to reduce shader register pressure. #6468
• [ET-VK] Removing tile input storage variable in conv_pw op and fetching the data in main loop. Also unrolling the main loop for performance improvement. #6467
• [ET-VK] Reduced int precision for texture coordinates in conv2d_pw op, to reduce shader register pressure and slightly improve performance. #6466
• -> [ET-VK] Reduced int precision for all int storage in q linear op, and reducing some texture coordinate storage variables to improve performance. #6465

This diff makes changes to the q linear op in the Vulkan backend of Executorch.
The changes includes reducing the precision of all int storage to reduce register usage and improve performance.
The diff also replaces some big persistent texture coordinate storage variables with smaller variables to further improve performance.

Differential Revision: D64780578

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6465

• 📄 Preview Python docs built from this PR

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:

• [PRE-EMPTIVE] Experimenting with new runners linux.aws.a100 on inductor-perf-compare.yml

✅ No Failures

As of commit 160b83d with merge base fa30e80 ():
💚 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]

This pull request was exported from Phabricator. Differential Revision: D64780578