[MITgcm-support] running mitcgm on Mac pro with openmpi
lian at ashimaresearch.com
Sun Jan 30 19:54:27 EST 2011
You method actually works. The problem was caused by
the lasted MITgcm build. I modified quite a bit of source codes
to accommodate my model, which worked fine with the older
version of MITgcm but not with the new ones (hard to keep up with
the updates:-) ).
I just grabbed the working MITgcm from cluster so everything works
just fine now. Thanks again for your help!
On 1/29/11 2:43 PM, Yuan Lian wrote:
> Hi Alastair,
> Thanks much! I just tried the openmpi from macports but
> the same problem still persists. I am wondering if this is
> caused by the 64 bit Snow Leopard or 64 bit compilers
> I am using. I will switch the system to 32bit and see how
> it goes.
> On 1/29/11 5:24 AM, Alastair McKinstry wrote:
>> On 2011-01-29 01:26, Yuan Lian wrote:
>>> Does anyone successfully run MITgcm on an eight-core Mac Pro?
>>> I am using Snow Leopard 64 bit, gfortran/gcc 4.6 and openmpi 1.5.1.
>>> I can compile the code without any trouble but
>>> 'mpirun -np 6 ./mitgcmuv' reports error with segmentation fault
>>> that is associated with 'libSystem.B.dylib'.
>> Try using 'mpiexec' rather than mpirun.
>> Attached is an optfile I use with a MacPro, gfortran and openmpi
>> It uses macports, so the paths might be slightly different, but worth
>>> Maybe someone here can help? Thanks!
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the MITgcm-support