Repository navigation
Patch pytensor on macos - #14
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
As I discovered during the RDkit UGM workshop the library is broken when pip-installed on MacOS and the native clang++ shipping with Xcode 27 or newer is used.
Compilation fails and disabling the compilation leads to very long execution time.
The proposed fix solves the upstream PyTensor issue (pymc-devs/pytensor#2268) by patching the flags passed to the
clang++.The devs apparently had some issues with new macOS linker - I have no experience to judge this, but omitting
-ld64flag works fine on my machine and all the testes are green.Alternative solution will be to use conda environment with
compilersinstalled but I guess most people will prefer to use normal venvs anyway.Also in my test patched version in venv is running tests faster then the same tests in the conda env (70s vs 90s).