[ET-VK][ez] Fix NaN propagation in binary div operations for padded texels

[SS-JIA]

Stack from ghstack (oldest at bottom):

• [ET-VK][ez] Restrict batch norm operator to 4D input tensors only #16420
• -> [ET-VK][ez] Fix NaN propagation in binary div operations for padded texels #16419

When the packed dimension size is not a multiple of 4, texture-backed
tensors have padding elements in the last texel. For division operations,
these padding regions contain 0/0 = NaN, which propagates through
subsequent reduction operations and corrupts results.

This fix adds conditional padding masking logic to binary_op shaders:

• Introduced MASK_PADDING codegen variable to binary_op.yaml
• Enabled MASK_PADDING=1 for binary_div and binary_floor_divide ops
• Added GLSL preprocessor macro definition in binary_op.glsl
• Implemented padding masking logic using modulo arithmetic to correctly
  identify last texels in batch concatenation scenarios
• Padding elements are explicitly zeroed out to prevent NaN propagation

The implementation follows GLSL best practices by using Python
preprocessing only for macro definition, keeping core shader logic
as pure GLSL with standard #ifdef directives.

Differential Revision: D89935220

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16419

• 📄 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:

• B200 runners are down due to network issues

❌ 2 New Failures, 1 Unrelated Failure

As of commit da3a2f1 with merge base 1066e7c ():

NEW FAILURES - The following jobs have failed:

• pull / test-samsung-models-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 4aaa33b2997f53db062355748920cc264252605f3ade8310f5ab9b4da6928472 /exec failed with exit code 1
• pull / test-samsung-quantmodels-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 36cfafb40c37362d39e4c92b1d5414c40ee70728ab25d393c5d404a2a73ca219 /exec failed with exit code 1

UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:

• pull / android / run-emulator (gh) (#16137)
  Timeout waiting for emulator to boot.

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

[github-actions]

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.

[mergennachin]

Can you add this regression test somewhere?

something like this:

x = torch.ones(2, 5, 3, 3)
y = torch.ones(2, 5, 3, 3)
return (x / y).sum()

We should at least add in vulkan test, but also, I wonder if we can expand it to Backend Test harness test that @GregoryComer built

[mergennachin]

const int texels_per_batch = divup4(out_sizes[packed_dim]); // -> this could be 0, and you're dividing by zero on the next line

i suggest a safer code:

    const int nspill = mod4(out_sizes[packed_dim]);
    if (nspill > 0) {
      const int texels_per_batch = divup4(out_sizes[packed_dim]); // this won't be 0 since we check nspill >0
      const bool is_last_texel = (lpos[packed_dim] % texels_per_batch) == (texels_per_batch - 1);
      if (is_last_texel) {
        [[unroll]] for (int i = nspill; i < 4; i++) {
          out_texel[i] = 0;
        }
      }
    }