VampirTrace is an open source library that allows detailed
logging of program execution for parallel applications using
message passing (MPI) and threads (OpenMP), Pthreads). Besides these typical
parallelization paradigms, VampirTrace is capable of tracing
GPU accelerated applications and generates exact time stamps
for all GPU related events.
Additional highlights are I/O event and memory allocation
tracing as well as recording of counter metrics like PAPI and external counters. To reduce the burden of the
underlying file system VampirTrace supports the I/O
Forwarding Scalability Layer (IOFSL), which allows users to write the data of
many trace files into one or a few physical files.
VampirTrace provides several variants to instrument the
application source code, such as manual, automatic by the
compiler, binary (Dyninst) instrumentation and library
wrapping.
The log data is written in the Open Trace Format (OTF), which can be analyzed and visualized by
the visualization tool Vampir.
VampirTrace has established a wide user base and has become an
integral part of Open MPI version 1.3 and higher.
Latest stable release:
A list of all important changes can be found in the
file ChangeLog.
Previous releases:
Older stable releases are still available in the archive.
Further links:
Please, report bugs, wishes, and suggestions to 
vampirsupport@zih.tu-dresden.de.
If you want to receive news about VampirTrace such as new
releases or major new features, please subscribe to the
Mailinglist.
VampirTrace is developed in collaboration with the KOJAK project at JSC/FZ Jülich.
Staff at the ZIH:
- Matthias Jurenz
- Dr. Andreas Knüpfer
- Ronny Brendel
- Matthias Lieber
- Jens Doleschal
- Joachim Protze
- Holger Mickler
- Michael Heyde
- Robert Dietrich
- Robert Schöne
- Dr. Matthias Müller
servicedesk@tu-dresden.de