Choose timezone
Your profile timezone:
MCGenerators review and planning meeting
→
Europe/Zurich
32/1-A24 (CERN)
Description
Title: Generator Services
Description: Generator Services yearly review meeting
Community: WLCG
Meeting Access Information:
- Meeting URL
http://evo.caltech.edu/evoNext/koala.jnlp?meeting=MsMiMI2s2vD8Dv9999Dt9s
- Phone Bridge
ID: 505 6813
Generator Services planning meeting, 25.04.2012
Present: W. Pokorski, G. Cosmo, M. Kirsanov, D.Tlisov
On EVO: T. Sjostrand, P. Richardson, G. Corti, L. Garren, D. Grellscheid, A. Buckley, J. Katzy, E. Lobodzinska, P. Loch
Status of Generator Services has been presented by W. Pokorski
Discussion on open questions:
1) Builds of generators with HepMC 2.03.11 still needed by ATLAS (and only ATLAS). Torbjorn pointed out again that this is quite some hassle to maintain two versions of the code (with IFDEFs) needed by the two versions of HepMC. ATLAS agreed to set the milestone of definitely freezing MCGenerators_hepmc2.03.11 by the end of 2012. After that date, no new versions of any generators will be build against HepMC 2.03.11 and therefore there will be no more the requirement for the generators code to compile against that version of HepMC.
2) Builds of generators with HepMC 2.06 series should follow LCGCMT cycle. The version of HepMC in LCGCMT is the one actually used by the experiments. In addition to that, Gloria has suggested to have builds with the latest version of HepMC, in order to allow testing.
3) MacOSX is actually not used by any of the experiments in the production. There is no real need to have full GENSER on MacOSX. We will therefore build only the main (new C++) generators on MacOSX, which come with the build systems working on this platform.
4) RPATH issues needs to be further discussed. It is clear that removing the RPATHs from the libraries would make them more portable, but we need to make sure that we don't break any existing functionality. Without RPATHs in GENSER libraries, users will be required to always set the different paths manually, in the environment variables. Witek will further talk to Benedikt Hegner to understand the situation better. The script from Anton K. for removing RPATHs need to be looked at closer before we decide to use it for GENSER. Witek will send this script to the interested people for evaluation.
5) ATLAS would be interested to have 'fake' debug libraries, i.e. <platform>_dbg links pointing to non-debug libraries. Witek has said that this is not done for any of the LCG external software, and that the 'fake' debug libraries are handled and the level of CMT interface packages. This would be the preferable solution from the point of view of GENSER. ATLAS will further investigate it.
6) David G. has pointed out that Grid UI does not contain CMAKE (the right version of it?) making it impossible to build, for instance, HepMC from sources. Witek will follow this up with the responsible people.
Present: W. Pokorski, G. Cosmo, M. Kirsanov, D.Tlisov
On EVO: T. Sjostrand, P. Richardson, G. Corti, L. Garren, D. Grellscheid, A. Buckley, J. Katzy, E. Lobodzinska, P. Loch
Status of Generator Services has been presented by W. Pokorski
Discussion on open questions:
1) Builds of generators with HepMC 2.03.11 still needed by ATLAS (and only ATLAS). Torbjorn pointed out again that this is quite some hassle to maintain two versions of the code (with IFDEFs) needed by the two versions of HepMC. ATLAS agreed to set the milestone of definitely freezing MCGenerators_hepmc2.03.11 by the end of 2012. After that date, no new versions of any generators will be build against HepMC 2.03.11 and therefore there will be no more the requirement for the generators code to compile against that version of HepMC.
2) Builds of generators with HepMC 2.06 series should follow LCGCMT cycle. The version of HepMC in LCGCMT is the one actually used by the experiments. In addition to that, Gloria has suggested to have builds with the latest version of HepMC, in order to allow testing.
3) MacOSX is actually not used by any of the experiments in the production. There is no real need to have full GENSER on MacOSX. We will therefore build only the main (new C++) generators on MacOSX, which come with the build systems working on this platform.
4) RPATH issues needs to be further discussed. It is clear that removing the RPATHs from the libraries would make them more portable, but we need to make sure that we don't break any existing functionality. Without RPATHs in GENSER libraries, users will be required to always set the different paths manually, in the environment variables. Witek will further talk to Benedikt Hegner to understand the situation better. The script from Anton K. for removing RPATHs need to be looked at closer before we decide to use it for GENSER. Witek will send this script to the interested people for evaluation.
5) ATLAS would be interested to have 'fake' debug libraries, i.e. <platform>_dbg links pointing to non-debug libraries. Witek has said that this is not done for any of the LCG external software, and that the 'fake' debug libraries are handled and the level of CMT interface packages. This would be the preferable solution from the point of view of GENSER. ATLAS will further investigate it.
6) David G. has pointed out that Grid UI does not contain CMAKE (the right version of it?) making it impossible to build, for instance, HepMC from sources. Witek will follow this up with the responsible people.
There are minutes attached to this event.
Show them.
