[MITgcm-support] Re: wacky mpi?
Ed Hill
ed at eh3.com
Sat Jun 19 21:44:58 EDT 2004
On Sat, 2004-06-19 at 16:03, N. Prive wrote:
> Hi,
>
> Not sure who is around these days, so I am mailing both of you!
>
> I'm having mpi problems on bradman. I changed three lines in my code,
> recompiled the gcm, and changed one number in a data file. When I try to
> run it, I get an error that I've never seen before, and the model doesn't
> even start to run. It says:
>
> Assertion failed: mpi_sgi_request, file req.c, line 1378
> IOT Trap
> MPI: MPI_COMM_WORLD rank 2 has terminated without calling MPI_Finalize()
> MPI: aborting job
> MPI: Receiver signal 6
>
> I checked with Will, and no settings have been changed on bradman since I
> ran the model on Thursday. Any idea what this error is an how I can fix
> it? I tried changing back the code to what I had when I ran the model
> last, and it didn't help.
Hi Nikki,
So even when you used the older (which had worked and was presumably
good) version of the code, it still gave the MPI error? That suggests
that it might not be an MITgcm problem per se -- perhaps its a library
or environment problem. Yes, I know thats vague but its hard to pin
down whats the matter without more detail.
So heres things you can do:
1) Do you have a simple MPI "hello world"-type program that you
can compile and run? That'll be a good test of the MPI bits.
eg.: http://www.parl.clemson.edu/~tdixon/tutorials/hello_world.html
2) Can you try running MITgcm without MPI? Does it work?
3) Wait until Monday and I'll come take a look.
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
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20040619/8494836a/attachment.sig>
More information about the MITgcm-support
mailing list