[Einstein Toolkit] #1471: Cactus should auto-detect newer versions of GCC from MacPorts

List overview All Threads
Download

newer

older

[Einstein Toolkit] #2144:...

[Einstein Toolkit] #2142: test

Einstein Toolkit

6 Nov 2013 6 Nov '13

2:47 p.m.

#1471: Cactus should auto-detect newer versions of GCC from MacPorts -------------------------+-------------------------------------------------- Reporter: hinder | Owner: Type: enhancement | Status: new Priority: minor | Milestone: Component: Cactus | Version: development version Keywords: | -------------------------+-------------------------------------------------- Currently, during Cactus configuration without an optionlist on Mac OS, I get

{{{ checking for gcc-mp-4.4... no checking for gcc-mp-4.3... no checking for gcc-mp-4.2... no checking for gcc... gcc checking whether the C compiler (gcc ) works... yes }}}

I have gcc-mp-4.6, and 4.7 and 4.8 are also available in macports. The configure script should be updated to detect these versions.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Show replies by date

Einstein Toolkit

20 Mar 20 Mar

9:11 a.m.

#1471: Cactus should auto-detect newer versions of GCC from MacPorts --------------------------+--------------------------------- Reporter: Ian Hinder | Owner: (none) Type: enhancement | Status: new Priority: minor | Milestone: Component: Cactus | Version: development version Resolution: | Keywords: --------------------------+---------------------------------

Comment (by Roland Haas):

Does MacPorts still call its compilers like this? If so then we should also add the homebrew equivalents ie gcc-6, gc-7 etc.

-- Ticket URL: http://cactuscode-20180221.cct.lsu.edu/trac/ticket/1471#comment:1 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

11:06 a.m.

Comment (by Steven R. Brandt):

Actually, I'm wondering if we need to bother with any of these checks. While "gcc" on Mac is clang, the ET compiles and the test suite runs (apart from the memory test thorn).

-- Ticket URL: http://cactuscode-20180221.cct.lsu.edu/trac/ticket/1471#comment:2 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

11:37 a.m.

Comment (by Roland Haas):

I guess the danger that would come to my mind would be if some of macports' packages are compiled with its own gcc or g++ and need a runtime library (eg C++ STL or something similar) that is not compatible with libraries that clang uses.

In your test code did you build some (all, a few) of the ExternalLibraries from the tarball in the ET or use them from macports (eg the ones that Ian's instructions ask to install)?

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:3 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

11:38 a.m.

Comment (by anonymous):

Good point, Roland. Not sure.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:4 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

27 Apr 27 Apr

10:01 a.m.

Comment (by Ian Hinder):

Steve, are you sure that the full ET can be compiled on a Mac with Apple's gcc? What does it use for Fortran compiler? It would be great to be able to compile without needing MacPorts' or HomeBrew's compilers installed, but I was never able to get this to work, and in the end gave up because I couldn't find a clean working solution better than what we currently have.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:5 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

10:07 a.m.

Comment (by Roland Haas):

It will use gfortran from HomeBrew or MacPorts (the later I don't know). So you end up with a mixed compile of clang+gfortran.

You still needs HomeBrew or MacPorts.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:6 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

10:26 a.m.

Comment (by Erik Schnetter):

Mixing Clang and GCC in this way leads to trouble. In particular, their OpenMP implementations are different, leading to link errors last time I checked. If linking worked, it is likely there would be segfaults at run time. You can make it work if you disable OpenMP.

I also don't see a good reason to mix compilers this way. If you install gfortran, you already have gcc and g++ as well, so you might as well use them to simplify things.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:7 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

30 Apr 30 Apr

9:54 a.m.

Comment (by Steven R. Brandt):

Sorry, I meant to reply to Ian's earlier comment and clarify.

I checked out a Mac from CCT, installed brew, then installed gcc, g++, gfortran, etc. (all the packages we recommend) then compiled the ET with generic.cfg and ran the test suite. Everything worked except for the Memory Test thorn.

I have not repeated this test for port as yet.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:8 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

10:28 a.m.

Comment (by Ian Hinder):

Do you know if it was using gcc from Homebrew, or if it was using the Apple-provided clang (named "gcc")? Regardless, if this is true, then it would simplify things for Homebrew users, if they didn't need to use a custom optionlist, and could instead use generic.cfg.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:9 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

Einstein Toolkit

10:30 a.m.

Comment (by Steven R. Brandt):

Ian, gcc -version showed that it was clang.

-- Ticket URL: https://trac.einsteintoolkit.org/ticket/1471#comment:10 Einstein Toolkit http://einsteintoolkit.org The Einstein Toolkit

ian.hinder＠aei.mpg.de

3:37 p.m.

...

On 30 Apr 2018, at 16:30, Einstein Toolkit trac-noreply@einsteintoolkit.org wrote:

#1471: Cactus should auto-detect newer versions of GCC from MacPorts --------------------------+--------------------------------- Reporter: Ian Hinder | Owner: (none) Type: enhancement | Status: new Priority: minor | Milestone: Component: Cactus | Version: development version Resolution: | Keywords: --------------------------+---------------------------------

Comment (by Steven R. Brandt):

Ian, gcc -version showed that it was clang.

And is this the compiler that Cactus picked up?

-- Ian Hinder http://members.aei.mpg.de/ianhin

3083

Age (days ago)

4719

Last active (days ago)

trac@lists.einsteintoolkit.org

11 comments

2 participants

tags (0)

participants (2)

Einstein Toolkit
ian.hinder＠aei.mpg.de