AF meeting 10th April 2014
Gabriele Cosmo, Gerri Ganis, Marco Cattaneo, Gulio Eulisse,
Pere Mato, Fons Rademakers, John Harvey, Peter Hristov, Nicolo Magini,
Federico Carminati, Andrea Valassi, Markus Elsing, Pedrag Buncic
Rolf Seuster, Marco Clemencic, Benedikt Hegner
- Action Items
- Line Management Issues (John)
- HEP SW Collaboration meeting feedback (all)
- Report from the LIM meeting (Benedikt)
- News from projects
- Experiments Feedback
1 (07/05/2014 @ 13:00)
- Setup a common calendar (Google?) with
the most important events. (John Harvey)
- Done. John will circulate the URL.
Management Issues (John)
- Technical Student started for the ROOT
project. He will be working migrating ROOT tests to CTest.
- The SFT fellow quota is oversubscribed
this year, so there will be no new requests.
HEP SW Collaboration meeting feedback
- More than 100 participants to the first meeting. The agenda
and materials are available.
- The summary was produced and we can add some corrections before
- The main agreement from the was to write a short white paper
describing the key characteristics of a HEP Software Foundation (goals,
scope, governance model, etc.). The proposed length is up to 5
pages and be available by May 12th.There the a long discussion during
the AF on how this first meeting was perceived.
- Here are some comments:
- We need a convergence of approaches
- Missing the goal or the scope. Collaboration responsible for
common tools. What about the software stack?
- Take the governance of what? We do not need to use the word
coordination (sharing, encouraging collaboration, ....)
- Has a problem with the 'chief architect'.
- The experiments were not involved in the preparation.
- There is no need for the software stack. Experiments will need to
build their own stack.
- Sharing the software. Defining ownership (license). Better
advertising. Encourage to have the means to support. Leave enough
room for innovation.
- Long term support. Enough interest from the Foundation may ensure
the long-term support.
- People has been reacting to fantasies (CERN centric, loss of
- Who are we? Ad personam? We can try to write something as AF
(all LHC experiments) and see if we agree.
- It was agreed among the AF to work on a single document expressing the
views of all the LHC experiments if possible. Peter volunteer to create
Google document for this.
from the LIM meeting (Benedikt)
- Provided new LCG releases (LCG 68 and
68root6) allowing for a ROOT 5 vs. ROOT 6 validation.
- LHCb discovered relocation-ability
issues in the ROOT builds introduced while moving to a new CMake based
build system for LCG
- Together with LHCb (many thanks for
the fruitful collaboration) set up a test infrastructure for rpm/apt
based SW distribution replacing the existing tarfile setup.
- This is the first time externals and
generators were provided in one go.
- Savannah to JIRA migration is ongoing
- ALICE migrations have been finished
- For the other experiments (all but
one) requested test migrations were provided.
- Some pending issues:
- There is one failing test migration
for the single biggest project "CMS Computing Infrastructure
Support". The import procedure overloads the server and this problem
could not be resolved yet.
- Existing programmatic interaction
with Savannah needs to be migrated to JIRA.
- All the LHCb projects to the test
instance are not visible through the firewall due to security
reasons. This is an annoyance, and is not making the testing
- Having contact people for ATLAS was
finally solved. The ATLAS computing projects final migrations to the
production instance have been started.
- Detailed progress can be followed up
- The SPI project got a summer student to
work on the Jenkins integration. If there are similar plans on the
experiments' side it would be useful to coordinate.
- Nothing to report from LIM as there was
no LIM since last AF.
- They have been cancelled for
collisions with other workshops.
- Version ROOT beta 3 was released
- c++11 is now enabled. Reduction of
- New technical student to re-work the
tests in CMake/CTest
- ROOT-6 news:
- The LLVM support or PCM is not going
to be ready for the scheduled May release. So, a sort of plan B was
put in place to emulate what PCM will bring to implement 'modular'
- Details were discussed at the ROOT 6
planning meetings with the experiments.
- Full stack build based on ROOT 6 beta
- Missing support Boost headers (using
instead C++11 equivalent functionality)
- Nothing in particular in Geant4.
- Work on the automatic garbage
collection (clearing the history)
- Migration to version 2.1 for servers
has started (Geant4 next week).
- New bug fix release of CernVM-FS
- CernVM 3.1.6 patch released with
OpenSSL bug fix.
- ROOT 6 migration team under the lead
of David Malon has been setup. The timeline is short. By September
they should be ready to be put in production.
- See the CORAL news.
- Two software stacks (with ROOT 5 and
ROOT 6) has been build to all the way to the analysis. The ROOT6 is
the default (if people do not specify it). Normal use cases are
- Geant4 10.p1 in the pre-release. It is
currently getting validated (a couple of days)
- MCJIT is a show-stopper for ARM and