Add a16w8 per-op test for var#19596

Merged

christine-long-meta merged 1 commit into

pytorch:mainfrom

christine-long-meta:export-D104532362

May 18, 2026

Contributor

christine-long-meta commented May 14, 2026 •

edited by meta-codesync Bot

Loading

Summary:
Add int16 activation / int8 weight (a16w8) quantization tests for aten.var on Ethos-U55 and Ethos-U85.

Changes

Add test_var_a16w8_u55_INT and test_var_a16w8_u85_INT using EthosU55PipelineINT/EthosU85PipelineINT with a16w8_quantization=True, symmetric_io_quantization=True
Split Var.test_parameters into passing (a16w8_var_test_parameters, keepdim=True) and xfail (a16w8_var_test_parameters_xfails, keepdim=False) groups for the a16w8 tests
Mark keepdim=False cases (var_3d_no_keep_dim_0_correction, var_4d_no_keep_dim_0_5_correction) as pytest.mark.xfail since var a16w8 produces incorrect output for scalar/reduced-rank output
Register ops/test_var.py in fbcode/ and xplat/ targets.bzl

Differential Revision: D104532362

christine-long-meta requested a review from digantdesai as a code owner

May 14, 2026 16:46

pytorch-bot Bot commented May 14, 2026 •

edited

Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19596

📄 Preview Python docs built from this PR

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

❗ 2 Active SEVs

There are 2 currently active SEVs. If your PR is affected, please view them below:

❌ 2 New Failures, 1 Unrelated Failure

As of commit c7709e1 with merge base 824cbff ():

NEW FAILURES - The following jobs have failed:

Build Presets / linux (llm, linux.2xlarge, executorch-ubuntu-22.04-clang12) / build (gh)
pull / test-multimodal-linux (gemma3-4b) / linux-job (gh)
RuntimeError: Command docker exec -t d23afe2ea7a3077d2fa3158f59e5bb2f5037481a7ddf89bb4af108d0b6fe869d /exec failed with exit code 139

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

pull / nxp-build-test / linux-job (gh) (matched linux rule in flaky-rules.json)
The process '/usr/bin/git' failed with exit code 128

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

meta-cla Bot added the CLA Signed label

github-actions Bot added ciflow/trunk module: arm labels

pytorch-bot Bot commented May 14, 2026 •

edited

Loading

~~Workflows were awaiting approval.~~ CI has now been triggered for the ciflow labels on this PR.

christine-long-meta force-pushed the export-D104532362 branch from 8989a2c to 6c91b26 Compare

May 14, 2026 16:47

meta-codesync Bot added fb-exported meta-exported labels

github-actions Bot commented May 14, 2026

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.

christine-long-meta force-pushed the export-D104532362 branch 2 times, most recently from 8989a2c to 6b40fe6 Compare

May 14, 2026 16:47

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var~~ Add a16w8 per-op test for var (#19596)

christine-long-meta force-pushed the export-D104532362 branch 4 times, most recently from d004872 to 578a5dd Compare

May 14, 2026 16:49

Contributor

meta-codesync Bot commented May 14, 2026

@christine-long-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D104532362.

christine-long-meta force-pushed the export-D104532362 branch 3 times, most recently from f1d1cf7 to 7aa22ac Compare

May 14, 2026 16:53

christine-long-meta added a commit to christine-long-meta/executorch that referenced this pull request


          Add a16w8 per-op test for var (pytorch#19596)

5e7cdc5

Summary:
Pull Request resolved: pytorch#19596

Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `a16w8_var_test_parameters` dict with 4 test configurations covering keepdim/no-keepdim and correction values 0, 0.5, and 1
- Add `test_var_a16w8_u55_INT` using `EthosU55PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True, qtol=128, epsilon=2**-16`
- Add `test_var_a16w8_u85_INT` using `EthosU85PipelineINT` with same kwargs
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

bypass-pytorch-oss-checks

Differential Revision: D104532362

christine-long-meta force-pushed the export-D104532362 branch from 7aa22ac to 5e7cdc5 Compare

May 14, 2026 16:53

Contributor

meta-codesync Bot commented May 14, 2026

@christine-long-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D104532362.

3 similar comments

Contributor

meta-codesync Bot commented May 14, 2026

@christine-long-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D104532362.

Contributor

meta-codesync Bot commented May 14, 2026

@christine-long-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D104532362.

Contributor

meta-codesync Bot commented May 14, 2026

@christine-long-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D104532362.

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var (#19596)~~ Add a16w8 per-op test for var

christine-long-meta force-pushed the export-D104532362 branch from 5e7cdc5 to 447a40d Compare

May 14, 2026 18:09

christine-long-meta added a commit to christine-long-meta/executorch that referenced this pull request


          Add a16w8 per-op test for var (pytorch#19596)

76e5af8

Summary:
Pull Request resolved: pytorch#19596

Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `a16w8_var_test_parameters` dict with 4 test configurations covering keepdim/no-keepdim and correction values 0, 0.5, and 1
- Add `test_var_a16w8_u55_INT` using `EthosU55PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True, qtol=128, epsilon=2**-16`
- Add `test_var_a16w8_u85_INT` using `EthosU85PipelineINT` with same kwargs
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

bypass-pytorch-oss-checks

Differential Revision: D104532362

christine-long-meta force-pushed the export-D104532362 branch from 447a40d to 76e5af8 Compare

May 14, 2026 20:51

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var~~ Add a16w8 per-op test for var (#19596)

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var (#19596)~~ Add a16w8 per-op test for var

christine-long-meta force-pushed the export-D104532362 branch from 76e5af8 to bd069fe Compare

May 15, 2026 19:02

christine-long-meta added a commit to christine-long-meta/executorch that referenced this pull request


          Add a16w8 per-op test for var (pytorch#19596)

f8f496d

Summary:
Pull Request resolved: pytorch#19596

Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `test_parameters_ethosu` class attribute to `Var` with 2 test configurations (4D tensors with correction=0 and correction=1)
- Switch existing `test_var_dim_u55_INT_no_dim` and `test_var_dim_u85_INT_no_dim` from `Var.test_parameters` to `Var.test_parameters_ethosu` for Ethos-U compatible tensor shapes
- Add `test_var_a16w8_u55_INT` using `EthosU55PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True`
- Add `test_var_a16w8_u85_INT` using `EthosU85PipelineINT` with same kwargs
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

Differential Revision: D104532362

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var~~ Add a16w8 per-op test for var (#19596)

christine-long-meta force-pushed the export-D104532362 branch from bd069fe to f8f496d Compare

May 15, 2026 19:25

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var (#19596)~~ Add a16w8 per-op test for var

christine-long-meta force-pushed the export-D104532362 branch from e2522ca to 7ef2930 Compare

May 16, 2026 01:09

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var~~ Add a16w8 per-op test for var (#19596)

christine-long-meta added a commit to christine-long-meta/executorch that referenced this pull request


          Add a16w8 per-op test for var (pytorch#19596)

e8a0b00

Summary:

Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `test_parameters_ethosu` class attribute to `Var` with 2 test configurations (4D tensors with correction=0 and correction=1)
- Switch existing `test_var_dim_u55_INT_no_dim` and `test_var_dim_u85_INT_no_dim` from `Var.test_parameters` to `Var.test_parameters_ethosu` for Ethos-U compatible tensor shapes
- Add `test_var_a16w8_u55_INT` using `EthosU55PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True`
- Add `test_var_a16w8_u85_INT` using `EthosU85PipelineINT` with same kwargs
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

Differential Revision: D104532362

christine-long-meta force-pushed the export-D104532362 branch from 7ef2930 to e8a0b00 Compare

May 16, 2026 02:04

christine-long-meta added a commit to christine-long-meta/executorch that referenced this pull request


          Add a16w8 per-op test for var (pytorch#19596)

8805beb

Summary:

Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `test_parameters_ethosu` class attribute to `Var` with 2 test configurations (4D tensors with correction=0 and correction=1)
- Switch existing `test_var_dim_u55_INT_no_dim` and `test_var_dim_u85_INT_no_dim` from `Var.test_parameters` to `Var.test_parameters_ethosu` for Ethos-U compatible tensor shapes
- Add `test_var_a16w8_u55_INT` using `EthosU55PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True`
- Add `test_var_a16w8_u85_INT` using `EthosU85PipelineINT` with same kwargs
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

Differential Revision: D104532362

christine-long-meta force-pushed the export-D104532362 branch from e8a0b00 to e861bdd Compare

May 16, 2026 02:07

christine-long-meta added a commit to christine-long-meta/executorch that referenced this pull request


          Add a16w8 per-op test for var (pytorch#19596)

e861bdd

Summary:

Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `test_parameters_ethosu` class attribute to `Var` with 2 test configurations (4D tensors with correction=0 and correction=1)
- Switch existing `test_var_dim_u55_INT_no_dim` and `test_var_dim_u85_INT_no_dim` from `Var.test_parameters` to `Var.test_parameters_ethosu` for Ethos-U compatible tensor shapes
- Add `test_var_a16w8_u55_INT` using `EthosU55PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True`
- Add `test_var_a16w8_u85_INT` using `EthosU85PipelineINT` with same kwargs
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

Differential Revision: D104532362

christine-long-meta force-pushed the export-D104532362 branch 2 times, most recently from 8805beb to 3879789 Compare

May 16, 2026 02:08

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var (#19596)~~ Add a16w8 per-op test for var

christine-long-meta force-pushed the export-D104532362 branch 2 times, most recently from 74f3654 to 0c1b2b7 Compare

May 17, 2026 16:06

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var~~ Add a16w8 per-op test for var (#19596)


          Add a16w8 per-op test for var

c7709e1

Summary:
Add int16 activation / int8 weight (a16w8) quantization tests for `aten.var` on Ethos-U55 and Ethos-U85.

## Changes
- Add `test_var_a16w8_u55_INT` and `test_var_a16w8_u85_INT` using `EthosU55PipelineINT`/`EthosU85PipelineINT` with `a16w8_quantization=True, symmetric_io_quantization=True`
- Split `Var.test_parameters` into passing (`a16w8_var_test_parameters`, keepdim=True) and xfail (`a16w8_var_test_parameters_xfails`, keepdim=False) groups for the a16w8 tests
- Mark `keepdim=False` cases (`var_3d_no_keep_dim_0_correction`, `var_4d_no_keep_dim_0_5_correction`) as `pytest.mark.xfail` since var a16w8 produces incorrect output for scalar/reduced-rank output
- Register `ops/test_var.py` in `fbcode/` and `xplat/` `targets.bzl`

Differential Revision: D104532362

meta-codesync Bot changed the title ~~Add a16w8 per-op test for var (#19596)~~ Add a16w8 per-op test for var

christine-long-meta force-pushed the export-D104532362 branch from 0c1b2b7 to c7709e1 Compare

May 17, 2026 22:52

Collaborator

zingo commented May 18, 2026

You seem to got a strange error in test-arm-ootb-linux, It seems very unrelated, I rerun the tests maybe it was a network glitch and will go away.

zingo approved these changes

View reviewed changes

Collaborator

zingo left a comment

The test re-run passed, approving :) Thanks!

christine-long-meta merged commit a142873 into pytorch:main

591 of 600 checks passed

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

Labels

ciflow/trunk CLA Signed fb-exported meta-exported module: arm