Choose timezone
Your profile timezone:
ROOT Review planning
ROOT planning 14 August
Present: The ROOT team, Benedikt, Peter, Liz, Bill
Actions:
o Axel: circulate a mail with the errors related to merging of pcms
Axel: News
o Can generate tiered modules in a stand alone example
o Can generate the tiered pcms for ROOT "by hand"
- Errors to be solved still there
- Will circulate a mail today with the erros to be fixed
o Optimisation level: decided not to "fix" clang. Fix expected by the LLVM Team.
o Fons working on exposure of clang symbols. Pere asks if working on the build
system would make sense knowing that ROOT may have a different symbols exposed.
This has no influence on the build system.
o Liz: is the optimisation bug a show stopper? No and is related to the interpreter
(even if not fixed by november, experiments will not be affected by this)
[Added in the minutes a posteriori: a symptom of the problem is the reduced speed of
execution of some macros in the interpreter, notably the ones generated when saving
a Canvas in .C format]
Danilo: Genreflex
o Working on propagating the attributes specified in selection rules to the reflection
information within ROOT (e.g. transient members, custom properties of classes like "id"
and so on)
o Gaudi+ROOT6: bug reported last week about Python.h header fixed, now lcg nightly
builds stuck with genconf (where python and c++ words meet). Most probably we need
to adapt genconf to the new plugin manager of Marco.
Peter Alice:
o All detectors are working. Still problems in the efficiency of matching. Despite
the fix of Andrei, still something is not the same between 5 and 6. We have to
wait for Andrei to be back.
o Ruben and Philippe: optimise baskets. With parameter tuning, big usage of memory
disappeared both in ROOT6 and 5.
Philippe:
o Code generation: supports templated functions.
o Prevent memory wastage with TFormula when using C++ functions
Bill:
o Waiting for Paul to finish Root 6 work and get back to CMS
Present: The ROOT team, Benedikt, Peter, Liz, Bill
Actions:
o Axel: circulate a mail with the errors related to merging of pcms
Axel: News
o Can generate tiered modules in a stand alone example
o Can generate the tiered pcms for ROOT "by hand"
- Errors to be solved still there
- Will circulate a mail today with the erros to be fixed
o Optimisation level: decided not to "fix" clang. Fix expected by the LLVM Team.
o Fons working on exposure of clang symbols. Pere asks if working on the build
system would make sense knowing that ROOT may have a different symbols exposed.
This has no influence on the build system.
o Liz: is the optimisation bug a show stopper? No and is related to the interpreter
(even if not fixed by november, experiments will not be affected by this)
[Added in the minutes a posteriori: a symptom of the problem is the reduced speed of
execution of some macros in the interpreter, notably the ones generated when saving
a Canvas in .C format]
Danilo: Genreflex
o Working on propagating the attributes specified in selection rules to the reflection
information within ROOT (e.g. transient members, custom properties of classes like "id"
and so on)
o Gaudi+ROOT6: bug reported last week about Python.h header fixed, now lcg nightly
builds stuck with genconf (where python and c++ words meet). Most probably we need
to adapt genconf to the new plugin manager of Marco.
Peter Alice:
o All detectors are working. Still problems in the efficiency of matching. Despite
the fix of Andrei, still something is not the same between 5 and 6. We have to
wait for Andrei to be back.
o Ruben and Philippe: optimise baskets. With parameter tuning, big usage of memory
disappeared both in ROOT6 and 5.
Philippe:
o Code generation: supports templated functions.
o Prevent memory wastage with TFormula when using C++ functions
Bill:
o Waiting for Paul to finish Root 6 work and get back to CMS
There are minutes attached to this event.
Show them.