Skip to content

[AutoDiff] Simplify AD-related SILGen logic. #26430

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 31, 2019

Conversation

dan-zheng
Copy link
Contributor

  • Do not unset @differentiable attribute functions for consistent
    same-module and cross-module [differentiable] SILGen behavior.
  • Always create vtable entry thunks for JVPs/VJPs.
    • Previously, vtable entry thunks were conditionally generated
      based on an ad-hoc condition.
  • Update tests.

Todo: remove JVP/VJP names from [differentiable] attribute.

- Do not unset `@differentiable` attribute functions for consistent
  same-module and cross-module `[differentiable]` SILGen behavior.
- Always create vtable entry thunks for JVPs/VJPs.
  - Previously, vtable entry thunks were conditionally generated
    based on an ad-hoc condition.
- Update tests.

Todo: remove JVP/VJP names from `[differentiable]` attribute.
@dan-zheng dan-zheng added the tensorflow This is for "tensorflow" branch PRs. label Jul 31, 2019
@dan-zheng dan-zheng requested a review from rxwei July 31, 2019 07:06
@dan-zheng
Copy link
Contributor Author

@swift-ci Please test tensorflow

@rxwei rxwei merged commit 79cdb96 into swiftlang:tensorflow Jul 31, 2019
@dan-zheng dan-zheng deleted the autodiff-silgen branch July 31, 2019 14:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
tensorflow This is for "tensorflow" branch PRs.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants