[petsc-users] mumps get stuck with parmetis

[Aron Ahmadia]

/project/k121/sandbox/petsc
-dev/externalpackages/MUMPS_4.10.0-p3/src/mumps_part9.F:4666
/project/k121/sandbox/petsc
-dev/externalpackages/MUMPS_4.10.0-p3/src/dmumps_part5.F:465
/project/k121/sandbox/petsc
-dev/externalpackages/MUMPS_4.10.0-p3/src/dmumps_part1.F:409
/project/k121/sandbox/petsc
-dev/externalpackages/MUMPS_4.10.0-p3/src/dmumps_part3.F:6651
/project/k121/sandbox/petsc
-dev/externalpackages/MUMPS_4.10.0-p3/src/mumps_c.c:422

I don't know the MUMPS source code very well so I couldn't tell you what
this set of routines are doing, but this is a snippet of the stack trace I
was seeing when the jobs died on BG/P.

If you set the "-info" flag on a PETSc run, it sends a lot of debugging
output to the screen, which is useful when you're in a situation where it
is hard to get access to a debugger or the stack trace.

A

On Tue, Apr 24, 2012 at 10:22 PM, Alexander Grayver <agrayver at gfz-potsdam.de
> wrote:

> Aron,
>
> This parameter let's to see mumps output in console. The important this to
> understand where mumps hangs, during analysis, factorization or actual
> solution (substitutions)? I'm almost sure it's factorization step. I
> observe this pretty often with mumps compiled with petsc (whereas when
> mumps is used directly it's quite rare to come along with this problem).
>
> Regards,
> Alexander
>
> ----- Reply message -----
> From: "Aron Ahmadia" <aron.ahmadia at kaust.edu.sa>
> To: "PETSc users list" <petsc-users at mcs.anl.gov>
> Subject: [petsc-users] mumps get stuck with parmetis
> Date: Tue, Apr 24, 2012 21:13
>
>
> I'm not sure if this is related, but Parmetis+Mumps+PETSc 3.2 on
> BlueGene/P was causing similar behavior without even setting any options.
>  The only way I was able to get a direct solver going was by switching over
> to SuperLU.
>
> A
>
> On Tue, Apr 24, 2012 at 10:01 PM, Alexander Grayver <
> agrayver at gfz-potsdam.de> wrote:
>
>> Can you set:
>>
>> -mat_mumps_icntl_4 1
>>
>> And send mumps output?
>> Also do you use lu or ilu? How large is your matrix?
>>
>> Regards,
>> Alexander
>>
>>
>> ----- Reply message -----
>> From: "Wen Jiang" <jiangwen84 at gmail.com>
>> To: <petsc-users at mcs.anl.gov>
>> Subject: [petsc-users] mumps get stuck with parmetis
>> Date: Tue, Apr 24, 2012 19:43
>>
>>
>> Hi,
>>
>> My code will hang at the solving stage when I use mumps with the runtime
>> option -mat_mumps_icntl_28 2 -mat_mumps_icntl_29 2. If I remove these two
>> options, my code works fine. I am using PETSc 3.2 and configure it with --download-mumps=1
>> and --download-parmetis=1. Could anyone give me any hints? Thanks.
>>
>> Regards,
>> Wen
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120424/3efe54e8/attachment-0001.htm>