Skip to content

Conversation

trivedivivek
Copy link
Contributor

Summary:
This diff issues related to using buffers for quantized linear weights in the Executorch Vulkan backend.

The changes include:

  • Using uint for buffer weight storage instead of uint8 and packing / unpacking integer values accordingly.
  • Updating the shader parameters in QuantizedLinearQCSNW.cpp to include the correct sizes for the packed weights.

Differential Revision: D84870681

@trivedivivek trivedivivek requested a review from SS-JIA as a code owner October 21, 2025 15:24
@pytorch-bot
Copy link

pytorch-bot bot commented Oct 21, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15306

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

❌ 8 New Failures, 1 Pending, 2 Unrelated Failures

As of commit ca24759 with merge base 6e08aef (image):

NEW FAILURES - The following jobs have failed:

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

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

@meta-cla meta-cla bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Oct 21, 2025
@meta-codesync
Copy link

meta-codesync bot commented Oct 21, 2025

@trivedivivek has exported this pull request. If you are a Meta employee, you can view the originating Diff in D84870681.

@trivedivivek trivedivivek added the release notes: vulkan Changes to the Vulkan backend delegate label Oct 21, 2025
trivedivivek added a commit to trivedivivek/executorch that referenced this pull request Oct 21, 2025
…ch#15306)

Summary:

This diff issues related to using buffers for quantized linear weights in the Executorch Vulkan backend.

The changes include:
* Using uint for buffer weight storage instead of uint8 and packing / unpacking integer values accordingly.
* Updating the shader parameters in `QuantizedLinearQCSNW.cpp` to include the correct sizes for the packed weights.

Reviewed By: SS-JIA

Differential Revision: D84870681
trivedivivek added a commit to trivedivivek/executorch that referenced this pull request Oct 21, 2025
…ch#15306)

Summary:

This diff issues related to using buffers for quantized linear weights in the Executorch Vulkan backend.

The changes include:
* Using uint for buffer weight storage instead of uint8 and packing / unpacking integer values accordingly.
* Updating the shader parameters in `QuantizedLinearQCSNW.cpp` to include the correct sizes for the packed weights.

Reviewed By: SS-JIA

Differential Revision: D84870681
@trivedivivek trivedivivek force-pushed the export-D84870681 branch 2 times, most recently from 86e5fbb to 8b259c6 Compare October 22, 2025 15:28
trivedivivek added a commit to trivedivivek/executorch that referenced this pull request Oct 22, 2025
…ch#15306)

Summary:

This diff issues related to using buffers for quantized linear weights in the Executorch Vulkan backend.

The changes include:
* Using uint for buffer weight storage instead of uint8 and packing / unpacking integer values accordingly.
* Updating the shader parameters in `QuantizedLinearQCSNW.cpp` to include the correct sizes for the packed weights.

Reviewed By: SS-JIA

Differential Revision: D84870681
…ch#15306)

Summary:

This diff issues related to using buffers for quantized linear weights in the Executorch Vulkan backend.

The changes include:
* Using uint for buffer weight storage instead of uint8 and packing / unpacking integer values accordingly.
* Updating the shader parameters in `QuantizedLinearQCSNW.cpp` to include the correct sizes for the packed weights.

Reviewed By: SS-JIA

Differential Revision: D84870681
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 meta-exported release notes: vulkan Changes to the Vulkan backend delegate

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants