Hi,
I was able to build petsc master with home-brew gcc 6.3 and intel 17.0 under
macOS sierra. As Mark said, it is important to reinstall the entire home-brew
compiler + libs after upgrading macOS (and often Xcode).
I am able to read / write exodus files. I am attaching my configure command
lines:
with intel 17.0:
./configure ./configure \
--download-chaco=1 \
--download-exodusii=1 \
--download-hdf5=1 \
--download-hypre=1 \
--download-metis=1 \
--download-ml=1 \
--download-netcdf=1 \
--download-parmetis=1 \
--download-triangle=1 \
--with-blas-lapack-dir=$MKLROOT \
--with-debugging=1 \
--with-mpi-dir=$MPI_HOME \
--with-pic \
--with-shared-libraries=1 \
--with-vendor-compilers=intel \
--with-x11=1
with gcc:
./configure \
--download-exodusii=1 \
--download-chaco=1 \
--download-ctetgen=1 \
--download-hdf5=1 \
--download-hypre=1 \
--download-metis=1 \
--download-ml=1 \
--download-netcdf=1 \
--download-parmetis=1 \
--download-triangle=1 \
--download-yaml=1 \
--with-debugging=1 \
--with-shared-libraries=1 \
--with-x11=1
Blaise
On Jan 23, 2017, at 7:55 AM, Mark Adams
<[email protected]<mailto:[email protected]>> wrote:
And I trust you updated all system software (like gcc, NetCDF and ExodusII).
OSX upgrades are hell.
On Mon, Jan 23, 2017 at 12:36 AM, Matthew Knepley
<[email protected]<mailto:[email protected]>> wrote:
On Sun, Jan 22, 2017 at 11:18 PM, Fande Kong
<[email protected]<mailto:[email protected]>> wrote:
Thanks, Matt,
Clang does not have this issue. The code runs fine with clang.
Okay, it sounds like a gcc bug on Mac 10.6, or at least in the version you have.
Matt
Fande,
On Sun, Jan 22, 2017 at 8:03 PM, Matthew Knepley
<[email protected]<mailto:[email protected]>> wrote:
On Sun, Jan 22, 2017 at 8:40 PM, Fande Kong
<[email protected]<mailto:[email protected]>> wrote:
Thanks, Matt.
It is a weird bug.
Do we have an alternative solution to this? I was wondering whether it is
possible to read the ".exo" files without using the ExodusII. For example, can
we read the ".exo" files using the netcdf only?
Well, ExodusII is only a think layer on NetCDF, just like other wrappers are
thin layers on HDF5. It is
really NetCDF that is failing. Can you switch compilers and see if that helps?
Matt
Fande Kong,
On Sun, Jan 22, 2017 at 6:50 PM, Matthew Knepley
<[email protected]<mailto:[email protected]>> wrote:
On Sun, Jan 22, 2017 at 5:28 PM, Fande Kong
<[email protected]<mailto:[email protected]>> wrote:
On Sun, Jan 22, 2017 at 12:35 PM, Matthew Knepley
<[email protected]<mailto:[email protected]>> wrote:
On Sun, Jan 22, 2017 at 12:41 PM, Fande Kong
<[email protected]<mailto:[email protected]>> wrote:
On Sat, Jan 21, 2017 at 10:47 PM, Matthew Knepley
<[email protected]<mailto:[email protected]>> wrote:
On Sat, Jan 21, 2017 at 10:38 PM, Fande Kong
<[email protected]<mailto:[email protected]>> wrote:
Hi All,
I upgraded the OS system to macOS Sierra, and observed that PETSc can not read
the exodus file any more. The same code runs fine on macOS Capitan. I also
tested the function DMPlexCreateExodusFromFile() against different versions of
the GCC compiler such as GCC-5.4 and GCC-6, and neither of them work. I guess
this issue is related to the external package exodus, and PETSc might not pick
up the right enveriment variables for the exodus.
This issue can be reproduced using the following simple code:
1) This is just a standard check. Have you reconfigured so that you know
ExodusII was built with the same compilers and system libraries?
2) If so, can you get a stack trace with gdb or lldb?
0 libsystem_kernel.dylib 0x00007fffad8b8dda __pthread_kill + 10
1 libsystem_pthread.dylib 0x00007fffad9a4787 pthread_kill + 90
2 libsystem_c.dylib 0x00007fffad81e420 abort + 129
3 libpetsc.3.7.dylib 0x00000001100eb9ee PetscAbortErrorHandler +
506 (errstop.c:40)
4 libpetsc.3.7.dylib 0x00000001100e631d PetscError + 916
(err.c:379)
5 libpetsc.3.7.dylib 0x00000001100ed830 PetscSignalHandlerDefault
+ 1927 (signal.c:160)
6 libpetsc.3.7.dylib 0x00000001100ed088
PetscSignalHandler_Private(int) + 630 (signal.c:49)
7 libsystem_platform.dylib 0x00007fffad997bba _sigtramp + 26
8 ??? 0x000000011ea09370 initialPoolContent + 19008
9 libnetcdf.7.dylib 0x000000011228fc62 utf8proc_map + 210
(dutf8proc.c:543)
10 libnetcdf.7.dylib 0x000000011228fd0f utf8proc_NFC + 38
(dutf8proc.c:568)
11 libnetcdf.7.dylib 0x00000001122a7928 NC_findattr + 110
(attr.c:341)
12 libnetcdf.7.dylib 0x00000001122a7a4e NC_lookupattr + 119
(attr.c:384)
13 libnetcdf.7.dylib 0x00000001122a93ef NC3_get_att + 47
(attr.c:1138)
14 libnetcdf.7.dylib 0x0000000112286126 nc_get_att_float + 90
(dattget.c:192)
15 libpetsc.3.7.dylib 0x00000001117f3a5b ex_open_int + 171
(ex_open.c:259)
16 libpetsc.3.7.dylib 0x0000000110c36609
DMPlexCreateExodusFromFile + 781 (plexexodusii.c:43)
17 DMPlexCreateExodusFromFile 0x000000010fed4cfc main + 397
(DMPlexCreateExodusFromFile.cpp:24)
18 libdyld.dylib 0x00007fffad78a255 start + 1
This is a NetCDF error on ex_open_int(). My guess is that your NetCDF build is
old and when it calls the system DLL
you bomb. Can you do a completely new build, meaning either reclone PETSc
somewhere else, or delete the whole
$PETSC_DIR/$PETSC_ARCH/externalpackage directory and reconfigure/build?
Thanks,
Matt
Hi Matt,
Thanks for reply. I recloned PETSc (the old petsc folder is deleted
completely) and reconfigure. And still has the same issue. I also checked if
the binary is complied against any other netcdf. The binary is actually
complied against the right netcdf installed through PETSc.
You can see that this crash happens on the call to
int CPU_word_size = 0, IO_word_size = 0, exoid = -1;
float version;
exoid = ex_open(filename, EX_READ, &CPU_word_size, &IO_word_size, &version);
which means the fault is not in PETSc, but rather in ExodusII for your machine.
We could definitely
confirm this if you made a 5 line program that only called this, but I don't
see why it should be different.
I am not sure what to do, since I am not in control of anything about ExodusII.
Can you report this to
their dev team? It is strange since Blaise has not reported this, and I know he
uses it all the time.
Thanks,
Matt
LiviadeMacBook-Pro:partition livia$ otool -L DMPlexCreateExodusFromFile
DMPlexCreateExodusFromFile:
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libpetsc.3.7.dylib
(compatibility version 3.7.0, current version 3.7.5)
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libsuperlu_dist.5.dylib
(compatibility version 5.0.0, current version 5.1.3)
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libparmetis.dylib
(compatibility version 0.0.0, current version 0.0.0)
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libmetis.dylib (compatibility
version 0.0.0, current version 0.0.0)
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libnetcdf.7.dylib
(compatibility version 10.0.0, current version 10.0.0)
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libhdf5_hl.8.dylib
(compatibility version 9.0.0, current version 9.1.0)
/Users/livia/math/petsc/arch-darwin-cxx-debug/lib/libhdf5.8.dylib
(compatibility version 9.0.0, current version 9.1.0)
/opt/X11/lib/libX11.6.dylib (compatibility version 10.0.0, current version
10.0.0)
/Users/livia/math/mpich-3.2_install/lib/libmpifort.12.dylib (compatibility
version 14.0.0, current version 14.0.0)
/usr/local/opt/gcc@5/lib/gcc/5/libgfortran.3.dylib (compatibility version
4.0.0, current version 4.0.0)
/usr/local/opt/gcc@5/lib/gcc/5/libquadmath.0.dylib (compatibility version
1.0.0, current version 1.0.0)
/Users/livia/math/mpich-3.2_install/lib/libmpicxx.12.dylib (compatibility
version 14.0.0, current version 14.0.0)
/usr/local/opt/gcc@5/lib/gcc/5/libstdc++.6.dylib (compatibility version 7.0.0,
current version 7.21.0)
/usr/local/opt/gcc@5/lib/gcc/5/libgcc_s.1.dylib (compatibility version 1.0.0,
current version 1.0.0)
/Users/livia/math/mpich-3.2_install/lib/libmpi.12.dylib (compatibility version
14.0.0, current version 14.0.0)
/Users/livia/math/mpich-3.2_install/lib/libpmpi.12.dylib (compatibility version
14.0.0, current version 14.0.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version
1238.0.0)
/usr/local/lib/gcc/5/libgcc_s.1.dylib (compatibility version 1.0.0, current
version 1.0.0)
Matt
static char help[] = " create mesh from exodus.\n\n";
#include <petscdm.h>
#include <petsc/private/dmpleximpl.h>
#undef __FUNCT__
#define __FUNCT__ "main"
int main(int argc,char **argv)
{
char fineMeshFileName[2048];
DM dm;
MPI_Comm comm;
PetscBool flg;
PetscErrorCode ierr;
ierr = PetscInitialize(&argc,&argv,(char *)0,help);CHKERRQ(ierr);
comm = PETSC_COMM_WORLD;
ierr =
PetscOptionsGetString(NULL,NULL,"-file",fineMeshFileName,sizeof(fineMeshFileName),&flg);CHKERRQ(ierr);
if(!flg){
SETERRQ(comm,PETSC_ERR_ARG_NULL,"please specify a fine mesh file \n");
}
ierr = DMPlexCreateExodusFromFile( comm,fineMeshFileName, PETSC_FALSE,
&dm);CHKERRQ(ierr);
ierr = DMDestroy(&dm);CHKERRQ(ierr);
ierr = PetscFinalize();CHKERRQ(ierr);
}
LiviadeMacBook-Pro:partition livia$ ./DMPlexCreateExodusFromFile -file Tri3.exo
[0]PETSC ERROR:
------------------------------------------------------------------------
[0]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably
memory access out of range
[0]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[0]PETSC ERROR: or see
http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind
[0]PETSC ERROR: or try http://valgrind.org<http://valgrind.org/> on GNU/linux
and Apple Mac OS X to find memory corruption errors
[0]PETSC ERROR: likely location of problem given in stack below
[0]PETSC ERROR: --------------------- Stack Frames
------------------------------------
[0]PETSC ERROR: Note: The EXACT line numbers in the stack are not available,
[0]PETSC ERROR: INSTEAD the line number of the start of the function
[0]PETSC ERROR: is given.
[0]PETSC ERROR: [0] DMPlexCreateExodusFromFile line 38
/Users/livia/math/petsc/src/dm/impls/plex/plexexodusii.c
[0]PETSC ERROR: --------------------- Error Message
--------------------------------------------------------------
[0]PETSC ERROR: Signal received
[0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for
trouble shooting.
[0]PETSC ERROR: Petsc Release Version 3.7.5, unknown
[0]PETSC ERROR: ./DMPlexCreateExodusFromFile on a arch-darwin-cxx-debug named
LiviadeMacBook-Pro.local by livia Sat Jan 21 21:04:22 2017
[0]PETSC ERROR: Configure options --with-clanguage=cxx
--with-shared-libraries=1 --download-fblaslapack=1 --with-mpi=1
--download-parmetis=1 --download-metis=1 --download-netcdf=1
--download-exodusii=1 --download-hdf5=1 --with-debugging=yes --with-c2html=0
--download-hypre=1 --with-64-bit-indices=1 --download-superlu_dist=1
PETSC_ARCH=arch-darwin-cxx-debug
[0]PETSC ERROR: #1 User provided function() line 0 in unknown file
application called MPI_Abort(MPI_COMM_WORLD, 59) - process 0
[unset]: write_line error; fd=-1 buf=:cmd=abort exitcode=59
:
system msg for write_line failure : Bad file descriptor
The log files of make and configuration are also attached. If you have any
idea on this issue, please let me know!
Fande Kong,
--
What most experimenters take for granted before they begin their experiments is
infinitely more interesting than any results to which their experiments lead.
-- Norbert Wiener
--
What most experimenters take for granted before they begin their experiments is
infinitely more interesting than any results to which their experiments lead.
-- Norbert Wiener
--
What most experimenters take for granted before they begin their experiments is
infinitely more interesting than any results to which their experiments lead.
-- Norbert Wiener
--
What most experimenters take for granted before they begin their experiments is
infinitely more interesting than any results to which their experiments lead.
-- Norbert Wiener
--
What most experimenters take for granted before they begin their experiments is
infinitely more interesting than any results to which their experiments lead.
-- Norbert Wiener
--
Department of Mathematics and Center for Computation & Technology
Louisiana State University, Baton Rouge, LA 70803, USA
Tel. +1 (225) 578 1612, Fax +1 (225) 578 4276 http://www.math.lsu.edu/~bourdin
