Make intdiv_256 a more generic binary operator template

[xuhdev]

Stack from ghstack:

• Make intdiv_256 a more generic binary operator template #35422 Make intdiv_256 a more generic binary operator template
• lshift and rshift on CUDA should match the behavior on CPU #35339 lshift and rshift on CUDA should match the behavior on CPU
• Utilize std::make_unsigned_t instead of wrapping lshift #35323 Utilize std::make_unsigned_t instead of wrapping lshift

This would make intdiv_256 a much more generic template that can easily accommodate other types of binary operators in the future. The operator becomes "out-of-place" because this would make it easier to substitute with other operators, and compilers should have no problem optimizing this.

Differential Revision: D20826861

[dr-ci]

💊 CircleCI build failures summary and remediations

As of commit 1db1672 (more details on the Dr. CI page):

---

💚 💚 Looks good so far! There are no CircleCI failures yet. 💚 💚

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker.

See how this bot performed.

This comment has been revised 2 times.

[ngimel]

Can you please comment on why are you making this change and why are you replacing an in-place version of the operation with the out-of-place?

[xuhdev]

Can you please comment on why are you making this change and why are you replacing an in-place version of the operation with the out-of-place?

Please see my update. I was doing this because I planned to add similar code for bit shifting operators, but it turned out to be unnecessary at this moment. Nevertheless, I decided to keep this part of the changes because it can be helpful for future incorporation of other binary operators.

[xuhdev]

@pytorchbot merge this please

[facebook-github-bot]

@ngimel merged this pull request in beac3f2.

[xuhdev]

This is actually merged in ea8021d