[MITgcm-support] Link problem on OSX

[Edward W Doddridge]

Hi Ruth,

That’s good to know. If I play with SIZE.h I can also make compilation fail with that error, but the problem goes away (for me at least) if I break the domain up into more tiles - I don’t know if that’s possible for you, but it might be an easy fix.

Ed


________________________________
Edward Doddridge
Postdoctoral researcher
Earth, Atmospheric and Planetary Sciences
Massachusetts Institute of Technology

www.doddridge.me<http://www.doddridge.me>

On 3 Jul 2019, at 14:37, Ruth Musgrave <rmusgrave at whoi.edu<mailto:rmusgrave at whoi.edu>> wrote:

Hi Ed,

I am able to compile, for example, the verification experiment exp4 at the default domain size. If I increase it to 200x200x100 grid points it fails with the “out of range” error as below. I’ve been able to run simulations this size (and much larger) on this laptop with no problems in the past.

Thanks,
Ruth

On Jul 3, 2019, at 2:01 PM, Edward W Doddridge <ewd at mit.edu<mailto:ewd at mit.edu>> wrote:

Hi Ruth,

Can you give some more details about the setup you’re trying to compile? Can you compile any of the verification experiments, or run `testreport`?

Ed


________________________________
Edward Doddridge
Postdoctoral researcher
Earth, Atmospheric and Planetary Sciences
Massachusetts Institute of Technology

www.doddridge.me<http://www.doddridge.me/>

On 3 Jul 2019, at 12:17, Ruth Musgrave <rmusgrave at whoi.edu<mailto:rmusgrave at whoi.edu>> wrote:

Hi

after some OS updates I suddenly found myself unable to compile MITgcm on my laptop. The initial issue was a problem with XCode, which I have resolved by upgrading my entire OS to Mojave, and doing all the updates. I am now running the latest, and undoubtably greatest, OSX.

However, I still cannot compile MITgcm. The issue seems to be in linking, with the following error in the last step:

ld: 32-bit RIP relative reference out of range (2154362563 max is +/-2GB): from _impldiff_ (0x10021DAF0) to _loctr.4787 (0x1808B09E0) in '_impldiff_' from impldiff.o for architecture x86_64

I have seen errors like this before and solved them with flags like -mcmodel=medium etc., but this time it doesn't work.

I followed with great interest this thread:

http://mailman.mitgcm.org/pipermail/mitgcm-support/2013-December/008687.html

but it is left with a cliffhanger! I did verify that my compilers are all 64 bit.

I would be very very grateful for any suggestions, as I am entirely out of ideas and desperate to run some simulations. I am using the build options in darwin_amd64_gfortran_netcdf.

Thank you
Ruth
_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org<mailto:MITgcm-support at mitgcm.org>
http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org<mailto:MITgcm-support at mitgcm.org>
http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org<mailto:MITgcm-support at mitgcm.org>
http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20190703/5dcd8e8a/attachment.html>