Dear Darshan Team,
I have installed the latest darshan version (3.3.0) for different MPIs
on our HPC JUWELS
(https://apps.fz-juelich.de/jsc/hps/juwels/configuration.html) and would
like to report two issues:
1. Intel (19.1.3.304) Compiler with IntelMPI/2019.8.254:
- Configure Step fails for the new APMPI feature:
/cd darshan-runtime; ./configure
--prefix=/path/to/darshan-runtime/3.3.0-iimpi-2020-APMPI
--with-mem-align=8 --with-log-path-by-env=DARSHAN_LOG_P//
//ATH --with-jobid-env=SLURM_JOBID CC=mpicc
--enable-hdf5-mod=$EBROOTHDF5 --enable-apmpi-mod --enable-apmpi-coll-sync/
- Error msg: /configure: error: APMPI module requires MPI version 3+/
- without the new APMPI Options the configure steps ends successfully:
/cd darshan-runtime; ./configure
--prefix=/p/software/juwels/stages/Devel-2020/software/darshan-runtime/3.3.0-iimpi-2020
--with-mem-align=8 --with-log-path-by-env=DARSHAN_LOG_PATH
--with-jobid-env=SLURM_JOBID CC=mpicc --enable-hdf5-mod=$EBROOTHDF5
/
2. GCC/9.3.0 Compiler with ParaStationMPI/5.4.7-1 (based on MPICH 3.3.2)
(https://github.com/ParaStation/psmpi/)
- darshan-runtime configured with /--enable-apmpi-mod
--enable-apmpi-coll-sync/
- For a simple helloworld code (MPI + OMP) the application seems to be
hanging in the MPI_FINALIZE call.
- if I open the *.darshan_partial file with `darshan-parser`, then the
following output is printed:
/Error: incompatible darshan file.//
//Error: expected version 3.21, but got //
//Error: failed to read darshan log file header.
/- There are no issues without APMPI.
3. Further questions:
- Is it possible to switch on/off APMPI during runtime?
- Are there any examples available that demonstrate the additional value
that can be achieved by using the new AutoPerf feature?
- Can you confirm that APMPI works on none-Cray systems ?
Best regards,
Thomas
--
Thomas Breuer
Division Application Support Forschungszentrum Jülich GmbH
Jülich Supercomputing Centre (JSC) Wilhelm-Johnen-Straße
http://www.fz-juelich.de/ias/jsc 52425 Jülich (Germany)
Phone: +49 2461 61-96742 (currently not available via phone)
Email: t.breuer(a)fz-juelich.de
-------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Volker Rieke
Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender),
Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt, Prof. Dr. Frauke Melchior
-------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------
Darshan users,
We are happy to announce a new pre-release for Darshan 3.3.0 (darshan-3.3.0-pre2). You can download the source HERE<ftp://ftp.mcs.anl.gov/pub/darshan/releases/darshan-3.3.0-pre2.tar.gz>.
This release contains a number of new features, bug fixes, and other improvements as detailed below:
* New PyDarshan Python package for analyzing Darshan log files
* PyDarshan provides a couple of interfaces to Darshan logs that should allow for easier development of custom Darshan log analysis utilities in Python
* See the PyDarshan documentation<https://www.mcs.anl.gov/research/projects/darshan/docs/pydarshan/index.html> for more details
* Thanks to Jakob Luettgau(DKRZ) for all of the hard work in contributing this package
* Bug fixes
* Modified Lustre module to use a safer method for obtaining Lustre file striping information (based on fgetxattr rather than ioctl)
* Fixed bug leading to potential deadlock when reducing shared records in MPI programs (known to affect mvapich2)
* Fixed bug causing errors when using Darshan's non-MPI mode when Darshan is built with an MPI compiler
* Disabled DXT's MPI-IO offset tracking for OpenMPI applications to avoid crashes caused by an OpenMPI bug
* Fixed various HDF5 module bugs:
* Fixes for applications using H5S_SELECT_NONE selections resulting in HDF5 error messages
* Fixes for applications using non-MPIIO VFDs resulting in HDF5 error messages
* Fixes for potentially incorrect counter values related to common accesses in the H5D module
* Other fixes allowing usage of the HDF5 modules in serial HDF5 applications
* Other enhancements
* Added support for querying Lustre file striping statistics for Lustre files that are symlinked from other file systems
* Added support for instrumenting openat, preadv, preadv2, pwritev, and pwritev2 functions, improving instrumentation of OpenMPI applications
* Improved error messages and documentation for darshan-util tools, including handling of incomplete Darshan log files
* Added new H5D module counter indicating the Darshan record ID of the file an HDF5 dataset belongs to
As always, please report any issues, comments, or questions to us using this mailing list or our GitLab page<https://xgitlab.cels.anl.gov/darshan/darshan>.
Thanks!
--Darshan team
