Skip to content

Conversation

@martinlsm
Copy link
Collaborator

@martinlsm martinlsm commented Oct 24, 2025

Move the insertion of INT8/INT32 RESCALE ops from the SUB node visitor to the pass InsertRescaleInt32Pass.

This is in practice a refactoring patch, but still the output TOSA files becomes different enough to cause Ethos-U55/U85 tests to fail in test_var.py and test_conv_relu_residual_add.py. However, the issue was fixed in https://gitlab.arm.com/artificial-intelligence/ethos-u/ethos-u-vela/-/commit/642f7517d3a6bd053032e1942822f6e38ccd546f so we temporarily set the failing tests to xfail until the version of Ethos-U Vela compiler depended on is bumped to one that includes the fix.

Test plan

test_insert_rescale_i32_pass.py has been modified to test the change.

cc @freddan80 @per @zingo @oscarandersson8218 @digantdesai

Martin Lindström and others added 2 commits October 24, 2025 14:14
Move the insertion of INT8/INT32 RESCALE ops from the ADD node visitor
to the pass InsertRescaleInt32Pass.

This is in practice a refactoring patch, but still the output TOSA file
becomes different enough to cause an Ethos-U55 test to fail in
test_conv_relu_residual_add.py. However, this issue was fixed in
https://gitlab.arm.com/artificial-intelligence/ethos-u/ethos-u-vela/-/commit/642f7517d3a6bd053032e1942822f6e38ccd546f
so we temporarily set the failing test to xfail until the version of
Ethos-U Vela compiler depended on is bumped to one that includes the
fix.

Signed-off-by: Martin Lindström <Martin.Lindstroem@arm.com>
Co-authored-by: Oscar Andersson <Oscar.Andersson@arm.com>
Change-Id: I90cd9ab5a296911a228b0080008e4c65ba773f7c
Move the insertion of INT8/INT32 RESCALE ops from the SUB node visitor
to the pass InsertRescaleInt32Pass.

This is in practice a refactoring patch, but still the output TOSA file
becomes different enough to cause an Ethos-U55 test to fail in
test_var.py. However, this issue was fixed in
https://gitlab.arm.com/artificial-intelligence/ethos-u/ethos-u-vela/-/commit/642f7517d3a6bd053032e1942822f6e38ccd546f
so we temporarily set the failing test to xfail until the version of
Ethos-U Vela compiler depended on is bumped to one that includes the
fix.

Signed-off-by: Martin Lindstroem <Martin.Lindstroem@arm.com>
Co-authored-by: Oscar Andersson <Oscar.Andersson@arm.com>
Change-Id: I38d63015e03e59c267338c84d64731b050854d06
@pytorch-bot
Copy link

pytorch-bot bot commented Oct 24, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15378

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

❌ 8 New Failures, 1 Cancelled Job, 3 Unrelated Failures

As of commit 07d8738 with merge base de56c81 (image):

NEW FAILURES - The following jobs have failed:

CANCELLED JOB - The following job was cancelled. Please retry:

FLAKY - The following job failed but was likely due to flakiness present on trunk:

BROKEN TRUNK - The following jobs failed but was 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 24, 2025
@martinlsm
Copy link
Collaborator Author

@pytorchbot label ciflow/trunk

@martinlsm
Copy link
Collaborator Author

@pytorchbot label "partner: arm"

@pytorch-bot pytorch-bot bot added the partner: arm For backend delegation, kernels, demo, etc. from the 3rd-party partner, Arm label Oct 24, 2025
@github-actions
Copy link

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ciflow/trunk CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. partner: arm For backend delegation, kernels, demo, etc. from the 3rd-party partner, Arm

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant