[PyTorch] Use real if constexpr behind macro in hot template

[swolchok]

Stack from ghstack:

• [PyTorch] Use real if constexpr behind macro in hot template #51368 [PyTorch] Use real if constexpr behind macro in hot template
• [PyTorch] Debug-gate static_assert in KernelFunction::makeFromUnboxedFunctor #51367 [PyTorch] Debug-gate static_assert in KernelFunction::makeFromUnboxedFunctor
• [PyTorch] Remove reference_cast in make_boxed_from_unboxed_functor #51319 [PyTorch] Remove reference_cast in make_boxed_from_unboxed_functor
• [PyTorch] Fix return value of IValue::to for Tensor/String #51463 [PyTorch] Fix return value of IValue::to for Tensor/String
• [PyTorch][easy] Fix TODOs in CppFunction constructors #51315 [PyTorch][easy] Fix TODOs in CppFunction constructors
• [PyTorch] Reduce template expansion in call_functor_with_args_from_stack #51313 [PyTorch] Reduce template expansion in call_functor_with_args_from_stack

This seems to noticably reduce build times, at least for
RegisterCPU.cpp. It makes sense that a compiler builtin would be
faster than simulating the same builtin with templates.

Identified with templight.

Differential Revision: D26154964

[facebook-github-bot]

💊 CI failures summary and remediations

As of commit f309519 (more details on the Dr. CI page):

---

• 1/1 failures possibly* introduced in this PR
  • 1/1 non-CircleCI failure(s)

---

Extra GitHub checks: 1 failed

• Failed: GitHub Actions - flake8-py3

---

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 to the (internal) Dr. CI Users group.

[bhosmer]

Do we have any data on the perf benefit from this? I can believe the lambdas in the polyfill API add overhead (or inline disruption risk, etc.) over the native feature, but given that we'll want to use this kind of surface-level #ifdef sparingly for readability's sake, it'd be good to establish a ballpark quantitative threshold for when it's justified.

[bhosmer]

Ah nvm my comment, saw the numbers in D26154964. To answer my own question, I'd say ~8% reduction in build time makes the surface-level splintering tolerable.

(BTW, wasn't marked as a reviewer on GH and missed it on Phab, hence the delay, apologies)

[swolchok]

@bhosmer note that IIUC, Ubuntu 16.04 LTS is the blocker for raising our minimum C++ version to C++17. It hits end of life in April, so we should be able to move to C++17 and use if constexpr directly in May.

[bhosmer]

@bhosmer note that IIUC, Ubuntu 16.04 LTS is the blocker for raising our minimum C++ version to C++17. It hits end of life in April, so we should be able to move to C++17 and use if constexpr directly in May.

Yup! @smessmer actually has the move to C++17 on his H1 roadmap, so you're right, it'll clean up pretty quick.

[facebook-github-bot]

This pull request has been merged in 6e1a5b1.