Don't use wrap_fx_proxy_cls for wrap_symint

[ezyang]

Stack from ghstack (oldest at bottom):

• Rename is_unspecialized to pass_arg_as_tensor, add comment #125496
• -> Don't use wrap_fx_proxy_cls for wrap_symint #125494
• Split wrap_symint out of wrap_unspecialized_primitive #125483
• Extend torch.utils._sympy.symbol for more Inductor symbols #125419

We use very little of the code in wrap_fx_proxy_cls, so dupe it out.

Signed-off-by: Edward Z. Yang ezyang@meta.com

cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/125494

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

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

✅ You can merge normally! (1 Unrelated Failure)

As of commit c1e817d with merge base 79af814 ():

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

• inductor / cuda12.1-py3.10-gcc9-sm86 / test (dynamic_inductor_timm, 2, 2, linux.g5.4xlarge.nvidia.gpu) (gh) (similar failure)
  sebotnet33ts_256

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

[lezcano]

Is this hotpath? Even if it is, I think it'd be better to keep both paths consistent, perhaps through an aux function

[ezyang]

I'm splitting them because I had a huge headache understanding the nest of conditionals for the two cases, and I'm about to add a third case (float) which is going to be subtly different in a few ways. So I'm deabstracting, working out the concrete details, and then we can reabstract when it's done.

[lezcano]

Splitting the madness that is wrap_fx_proxy_cls SGTM. We should make sure that we don't leave duped code at the end of this refactor tho.