-
-
Notifications
You must be signed in to change notification settings - Fork 503
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
The "-march=native" flag vs. the compiler used when Python extensions are built #31228
Comments
comment:2
Do we manually set the Compiler to be different somewhere vor is this enforced by using system python? If not, is there a reasonable place to check this? Can this happen to python modules only? |
comment:3
There is |
comment:4
This is the standard behavior of python's distutils. Only Python extension modules are affected.
These are probably the best places to put additional checks. |
comment:5
I see. I can add a file compiling with If this fails, I have to build extension modules with |
comment:6
Much easier would be to give completely up on |
comment:7
Replying to @kliem:
Yes, I think that would be better. Otherwise too many |
Commit: |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
Dependencies: #31132 |
comment:11
Let's do this on top of #31132, which has modified the same files. I have added one commit that refactors |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:13
Here's an attempt |
comment:14
Looks plausible, however I don't see exactly, where
Maybe we need to add
But I'm just hoping they are respected. |
comment:15
Yes, I meant to add |
Author: Matthias Koeppe |
New commits:
|
comment:17
On homebrew, this will have to be tested together with #31227 if a current XCode version is in use |
comment:18
There is a merge conflict with #31227. Does the problem show up during github workflows? (If so, I can fix the merge conflict and test whether the problem goes away.) |
comment:19
It showed up when I tried in on my local machine. |
comment:20
Instead of disabling |
comment:21
I prefer the easy solution as currently implemented, see comment 7 above |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:26
Upgrading this to blocker, because this ticket is needed to get cython installation to work again (when python uses clang instead of gcc): |
Reviewer: Jonathan Kliem, https://github.com/kliem/sage/pull/35/checks |
comment:29
I just started github actions along with commit Please tell me, if I need to pull other tickets yet as mentioned in: Btw, I personally would find it useful, those tickets needed to fix github actions on top of the current beta are collected somewhere. |
Changed author from Matthias Koeppe to none |
Changed branch from u/mkoeppe/the___march_native__flag_vs__the_compiler_used_when_python_extensions_are_built to none |
Changed reviewer from Jonathan Kliem, https://github.com/kliem/sage/pull/35/checks to Jonathan Kliem |
Changed commit from |
comment:33
Agreed. |
(from #31132)
#27122 determines whether
CC
accepts-march=native
and then adds it globally toCFLAGS
inbuild/bin/sage-build-env
.However, Python extension modules are compiled with the compiler listed in sysconfig. On macOS with homebrew, using
/usr/bin/python3
, this could beclang
, which does not accept-march=native
.Depends on #3113
Depends on #30725
CC: @kliem @zlscherr @jhpalmieri
Component: build: configure
Reviewer: Jonathan Kliem
Issue created by migration from https://trac.sagemath.org/ticket/31228
The text was updated successfully, but these errors were encountered: