Arm backend: Fuse duplicate user ops

[AdrianLundell]

Adds a pass which checks if a node has multiple users performing equivalent operations its output. If that is the case, it fuses these ops into one.

cc @freddan80 @per @zingo @oscarandersson8218 @digantdesai

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15218

• 📄 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:

• Long queue for ROCM runners, also B200 and XPU queueing is observed

❌ 1 New Failure, 2 Unrelated Failures

As of commit 11d76ab with merge base 94ce147 ():

NEW FAILURE - The following job has failed:

• trunk / test-llama-runner-mac (fp32, coreml) / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 1

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

• trunk / unittest-release / macos / macos-job (gh) (similar failure)
  [ FAILED ] WavLoaderTest.LoadAudioDataRejectsUnsupportedFormat

BROKEN TRUNK - The following job failed but was present on the merge base:

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

• trunk / unittest-release / linux / linux-job (gh) (trunk failure)
  [ FAILED ] WavLoaderTest.LoadAudioDataRejectsUnsupportedFormat

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

[zingo]

I'll rebase to get the test-mcu-model error out of the way

[zingo]

Failed with
FAILED backends/arm/test/ops/test_tanh.py::test_tanh_16a8w_tosa_INT[rand] - AssertionError: Output 0 does not match reference output.
Given atol: 0.0010152587890625, rtol: 0.001.
Output tensor shape: torch.Size([10, 10]), dtype: torch.float32
Difference: max: 1.52587890625e-05, abs: 0.00152587890625, mean abs error: 2.6092529296875e-05.
-- Model vs. Reference --
Numel: 100, 100
Median: -0.0678253173828125, -0.0678253173828125
Mean: -0.0142254638671875, -0.014200897216796874
Max: 0.4605865478515625, 0.4621124267578125
Min: -0.4584808349609375, -0.4584808349609375
= 1 failed, 3050 passed, 2534 skipped, 2239 xfailed, 115 xpassed, 982 warnings in 2753.46s (0:45:53) =

This is probably just a flakey test Retrigger

[zingo]

Added files in _passes is part of a buck2 *.py rule so this should work in buck2
Fails are unrelated