[MITgcm-support] bluesky build
Baylor Fox-Kemper
baylor at MIT.EDU
Tue Jun 7 13:49:10 EDT 2005
Hi Mike,
Are you using singlecpuio or useglobalfiles? Those can be very slow
in mpi on certain systems, I think. I can't remember now, but one of
those really slowed down some interactive jobs I was doing at GFDL.
Cheers,
-Baylor
On Jun 7, 2005, at 1:43 PM, Ed Hill wrote:
> On Tue, 2005-06-07 at 11:25 -0400, Mike Spall wrote:
>> Hi All,
>>
>> I have tried compiling MITgcm on ncar's bluesky. It seems
>> to be working in that I get no errors, but it also seems
>> to be taking a very long time. I compiled with
>> MITgcm/tools/build_options/sp4+mpi yesterday and
>> it was not done after > 20 minutes so I killed it.
>> I was doing this interactively, not by batch.
>> Does this sound right?
>
> Hi Mike,
>
> This is only a guess, but it sounds like the very slow build could be
> due to slow disk I/O. Please contact the bluesky admins since there is
> almost nothing that we can do about it from our side.
>
> Also, if you are doing a build on an SMP system and you have access to
> Gnu make ("gmake") or a similarly capable "make", then you can do the
> following:
>
> ../../../tools/genamke2 ...OTHER_OPTIONS...
> make depend
> make -j 4
>
> where the "-j 4" specifies that four jobs should be run in parallel.
> On
> SMP systems, this can result in huge speedups of the build. For
> instance, "-j 3" on dual-processor Opterons and "-j 16" on some Altix
> systems can provide massive (approximately, x2 and x14) speedups of the
> compilation stage.
>
> Ed
>
> --
> Edward H. Hill III, PhD
> office: MIT Dept. of EAPS; Rm 54-1424; 77 Massachusetts Ave.
> Cambridge, MA 02139-4307
> emails: eh3 at mit.edu ed at eh3.com
> URLs: http://web.mit.edu/eh3/ http://eh3.com/
> phone: 617-253-0098
> fax: 617-253-4464
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
More information about the MITgcm-support
mailing list