Minutes SB 51, December 9th 2008 v.1.0
Participants: M.Asai, H.Kurasige, P.Gumplinger, G.Cosmo, G.Folger,
S.Banijeree, V.Ivantchenko (until 9:30), J.Apostolakis, M.Verderi,
G.Cuttone, D.Wright, S.Incerti, J.Perl, T.Sasaki.
Minutes: G.Cosmo (9-DEC-2008)
41.2 - no news
42.3 - no news
44.1 - not yet discussed; people from space community willing to donate data
in G4NDL format.
44.2 - no news
44.3 - contacted but no news... closed
44.4 - Done. Close
45.2 - No news
46.1 - Done. Closed
46.3 - Not done yet.
47.1 - No news
47.3 - 66% done, one at 80% to be closed.
48.3 - Coupled to the US-Geant4 funding proposal. To be closed.
48.4 - Done, added to future plans for hadronics. Closed.
48.5 - Done. Closed.
49.1 - Resolved. Closed.
49.2 - Done. Closed.
49.3 - List of items ready, not clear for the fundings (more clear as of April 2009)
49.4 - Exists but not yet available from the pages
49.5 - Not done yet, planning for release 9.2.
John resumes the discussion done in the preparation meeting (impossible to
summarise it here, JA has the summary...). All agree in silence.
- First draft of "join-EM" web page
- Rec.3 (tradeoff between physics accuracy and speed): missing a benchmark for
- Rec.4 (guidance on choice of EM models): ...
- Rec.5 (single EM package): a lot of progress ...
- Rec.6 (Fluka interface): contacted Fluka people, no change from their side. ILC
seems to have given up on their request, too difficult and no manpower...
G.Cuttone proposes to discuss with Fluka people (ACTION-1).
- Rec.7 (plans, improvement to hadronics): ...
- Rec.8 (simple hadronic benchmarks): give a comprehensive view of what we
have already for hadronics in the existing testing suite. Make it available from
the web (ACTION-2).
- Rec.12 (performance optimization guide): two presentations exist. No physics or
too scattered information. Start to put together a Twiki page which summarises
everything, with a plan for having in the official documentation by May 2009.
- Rec.13 (turning off irrelevant processes):...
- Rec.14 (guidance on variance reduction techniques): resource limited, could put
together a simple page of information, eventually link it to the performance Twiki:
importance biasing, cross-section biasing, leading-particle biasing, ...
JA volunteer to summarise the different techniques existing starting from Alex
Howard presentation at Hebden-Bridge. (ACTION-4)
- Rec.15 (systematic tracking of code performance): ACTION-5 - identify responsible
in each Category for running benchmarks and provide progress every other
- Rec.17 (plans for performance): ...
- Rec.19 (benchmarking against other MC codes): systematically ? Perhaps it make
sense to compare to other MCs where there is no data available...
- Rec.23 (web update): JA take over responsibility for this (no longer Katsuya).
There has been improvements. Should avoid duplication. Up-to-date publications
lists made available in WG pages should be linked from a single source.
- Rec.24 (documenting limitation of physics models): ...
- Rec.25 (code documentation, LXR): LXR is OK. SRM will be kept as mean for
commenting code and define a policy. In addition we should add a link to the
existing Doxygen docs.
- Rec.26 (Release notes): progress made.
- Rec.27 (Installation, RPM): ...
- Rec.28 (Wiki): ...
- Rec.29 (user-driven physics lists): some examples in the past of commitment
taken by the Collaboration can be used as response for a positive answer on a
case-by-case approach, based on resources to be found for maintenance.
JA takes this on board.
- Rec.30 (man power consistent with plans): ...
- Rec.31 (additional man power): ...
- Bug in G4SteppingManager proposed by Makoto for release 9.2.
- Livermore-based extensions are not included in release 9.2, delayed by
developers, eventually available in the first half of 2009.
Plans for 2009 and beyond
- ACTION-6 (JA): Organise phone meeting (10-12 January) to discuss proposals.
- Identify required man power for long-term planning in order to address major
- Need to understand shortly how to prepare the long-term plans, web pages
Meeting closed at 11:00.