[CIG-LONG] PETSc error when using MUMPS

Luke Mondy luke.s.mondy at gmail.com
Wed Jun 23 01:12:23 PDT 2010


Hello,

I have been trying to compile Gale from the mercurial repos, and I have run
into PETSc error troubles when I try to run a multi-processor job with
MUMPS.

Here the processes I have gone through on a Ubuntu 10.04 x64 bit VM:
1) Apt-get'ed libxml2-dev, mpich2 (with the mpd daemon setup), mercurial

2) Downloaded and installed HDF5 (with no problems)

3) Downloaded petsc3.0.0-p12, and extracted it to /usr/local/petsc-3.0.0-p12

4) Chown'ed the whole directory and contents to my username

5) In the petsc folder, ran the following command:

config/configure.py --download-mumps=1 --download-parmetis=1
--download-scalapack=1 --download-f-blas-lapack=1 --download-blacs=1
--with-shared=1

(Just to note, right after the "Configuring PETSc.." heading, I get this:

/usr/local/petsc-3.0.0-p12/config/BuildSystem/config/compilers.py:7:
DeprecationWarning: the sets module is deprecated
  import sets
/usr/local/petsc-3.0.0-p12/config/PETSc/package.py:7: DeprecationWarning:
the md5 module is deprecated; use hashlib instead
  import md5
/usr/local/petsc-3.0.0-p12/config/BuildSystem/script.py:101:
DeprecationWarning: The popen2 module is deprecated.  Use the subprocess
module.
  import popen2

after which, the compiling and installing of all the requested modules is as
normal)

After all that, it says:

Compilers:
  C Compiler:         mpicc  -fPIC -Wall -Wwrite-strings
-Wno-strict-aliasing -g3
  Fortran Compiler:   mpif90  -fPIC -Wall -Wno-unused-variable -g
Linkers:
  Shared linker:   mpicc  -shared  -fPIC -Wall -Wwrite-strings
-Wno-strict-aliasing -g3
  Dynamic linker:   mpicc  -shared  -fPIC -Wall -Wwrite-strings
-Wno-strict-aliasing -g3
PETSc:
  **
  ** Before running "make" your PETSC_ARCH must be specified with:
  **  ** setenv PETSC_ARCH linux-gnu-c-debug (csh/tcsh)
  **  ** PETSC_ARCH=linux-gnu-c-debug; export PETSC_ARCH (sh/bash)
  **
  **
  ** Before running "make" your PETSC_DIR must be specified with:
  **  ** setenv PETSC_DIR /usr/local/petsc-3.0.0-p12 (csh/tcsh)
  **  ** PETSC_DIR=/usr/local/petsc-3.0.0-p12; export PETSC_DIR (sh/bash)
  **
  **
  ** Now build the libraries with "make all"
  **
  Clanguage: C
  Scalar type:real
MPI:
  Includes: -I/usr/include/mpich2
  PETSc shared libraries: enabled
  PETSc dynamic libraries: disabled
BLAS/LAPACK: -Wl,-rpath,/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
-L/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib -lflapack
-Wl,-rpath,/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
-L/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib -lfblas
ParMetis:
  Includes: -I/usr/include/mpich2
-I/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/include
  Library:  -Wl,-rpath,/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
-L/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib -lparmetis -lmetis
SCALAPACK:
  Library:  -Wl,-rpath,/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
-L/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib -lscalapack
blacs:
  Library:  -Wl,-rpath,/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
-L/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib -lblacs
MUMPS:
  Includes: -I/usr/include/mpich2
-I/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/include
  Library:  -Wl,-rpath,/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
-L/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib -lcmumps -ldmumps
-lsmumps -lzmumps -lmumps_common -lpord

6) Put these into my .bashrc

export PETSC_DIR=/usr/local/petsc-3.0.0-p12
export PETSC_ARCH=linux-gnu-c-debug

and then restarted my terminal.

7) Then I ran the "make all test" command, and at the end got this:

Running test examples to verify correct installation
C/C++ example src/snes/examples/tutorials/ex19 run successfully with 1 MPI
process
C/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 MPI
processes
Fortran example src/snes/examples/tutorials/ex5f run successfully with 1 MPI
process
Completed test examples

7) Mercurial'ed the Gale repo into my home directory.

8) I then ran ./configure.sh, and get:

earthbyte at earthbyte-virtualmachine:~/Programs/gale$ ./configure.sh scons:
Reading SConscript files ...
Checking whether the C compiler worksyes
Checking for package libXML2... yes
Checking for package MPI... yes
Checking for package HDF5... yes
Checking for package PETSc... yes
Checking for package PETScExt... no

C compiler:     'mpicc'
C flags:        -g -fPIC -Wwrite-strings -Wno-strict-aliasing -g3 -Wall
C preprocessor: ['DEBUG', 'HAVE_LIBXML2', 'HAVE_MPI', 'HAVE_HDF5',
'HAVE_PETSC']


***************************************
* Successfully configured.            *
* Now run './scons' to build.      *
***************************************

scons: done reading SConscript files.
scons: Building targets ...
scons: `.' is up to date.
scons: done building targets.

(A side note, is PETScExt needed? I have tried installing it, but it comes
up saying some odd things...)

9) I then ran ./scons -j 2, which eventually ends with:

scons: done building targets.

10) Then, if I run:

build/bin/Gale input/cookbook/yielding.xml -mat_type aijumps -ksp_type pre
only -pc_type lu

or:

mpirun -np 2 build/bin/Gale input/cookbook/yielding.xml -mat_type aijumps
-ksp_type pre only -pc_type lu

or:

build/bin/Gale input/cookbook/yielding.xml -ksp_type pre only -pc_type lu

I get major PETSc errors - here is a snapshot:

[0]PETSC ERROR: build/bin/Gale on a linux-gnu named earthbyte-virtualmachine
by earthbyte Wed Jun 23 18:09:11 2010
[0]PETSC ERROR: Libraries linked from
/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
[0]PETSC ERROR: Configure run at Wed Jun 23 17:21:05 2010
[0]PETSC ERROR: Configure options --download-mumps=1 --download-parmetis=1
--download-scalapack --download-f-blas-lapack=1 --download-blacs=1
--with-shared=1
[0]PETSC ERROR:
------------------------------------------------------------------------
[0]PETSC ERROR: MatSetValues() line 900 in src/mat/interface/matrix.c
[0]PETSC ERROR: --------------------- Error Message
------------------------------------
[0]PETSC ERROR: Object is in wrong state!
[0]PETSC ERROR: Mat object's type is not set: Argument # 1!
[0]PETSC ERROR:
------------------------------------------------------------------------
[0]PETSC ERROR: Petsc Release Version 3.0.0, Patch 12, Tue Mar 16 23:20:08
CDT 2010
[0]PETSC ERROR: See docs/changes/index.html for recent updates.
[0]PETSC ERROR: See docs/faq.html for hints about trouble shooting.
[0]PETSC ERROR: See docs/index.html for manual pages.
[0]PETSC ERROR:
------------------------------------------------------------------------

this showed up too:

[0]PETSC ERROR: [1]PETSC ERROR: --------------------- Error Message
------------------------------------
[1]PETSC ERROR: Unknown type. Check for miss-spelling or missing external
package needed for type
 seehttp://
www.mcs.anl.gov/petsc/petsc-as/documentation/installation.html#external!
[1]PETSC ERROR: Unknown Mat type given: aijumps!
[1]PETSC ERROR:
------------------------------------------------------------------------
[1]PETSC ERROR: Petsc Release Version 3.0.0, Patch 12, Tue Mar 16 23:20:08
CDT 2010
[1]PETSC ERROR: See docs/changes/index.html for recent updates.
[1]PETSC ERROR: See docs/faq.html for hints about trouble shooting.
[1]PETSC ERROR: See docs/index.html for manual pages.
[1]PETSC ERROR:
------------------------------------------------------------------------
[1]PETSC ERROR: build/bin/Gale on a linux-gnu named earthbyte-virtualmachine
by earthbyte Wed Jun 23 18:09:11 2010
[1]PETSC ERROR: Libraries linked from
/usr/local/petsc-3.0.0-p12/linux-gnu-c-debug/lib
[1]PETSC ERROR: Configure run at Wed Jun 23 17:21:05 2010
[1]PETSC ERROR: Configure options --download-mumps=1 --download-parmetis=1
--download-scalapack --download-f-blas-lapack=1 --download-blacs=1
--with-shared=1
[1]PETSC ERROR:
------------------------------------------------------------------------
[1]PETSC ERROR: MatSetType() line 52 in src/mat/interface/matreg.c
[1]PETSC ERROR: MatSetFromOptions() line 195 in src/mat/utils/gcreate.c
--------------------- Error Message ------------------------------------


Any ideas what is going wrong?

Kind Regards,

Luke Mondy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://geodynamics.org/pipermail/cig-long/attachments/20100623/3852fe3b/attachment.htm 


More information about the CIG-LONG mailing list