aszostak [Fri, 11 Nov 2011 14:17:44 +0000 (14:17 +0000)]
Will now forward the whole trigger meta data structure normally passed to HLT components as a data block through TCPDumpSubscriber. So updating datatype appropriately.
marian [Fri, 11 Nov 2011 10:18:22 +0000 (10:18 +0000)]
Making drift velocity calibration more robust
Stability of algorithm tested with standard run production (HVscan):
154793
153116
146859
149102
No single software failuere. The mean fraction of outlyers filtered ~ 5 %.
Code modifications:
1. Usage of the internal robust option in the AliRelAlegnerKalman.
Both median and nsignma rejection.
2. Increasing the histogram range for the dZ(6 cm) and dTheta(0.05) histogramming
pchrist [Thu, 10 Nov 2011 19:09:03 +0000 (19:09 +0000)]
Adding the possibility to analyze particles based on their pdg code (MC) + addition of IsPhysicalPrimary + changes in the steering and AddTask macro to be used for the MC analysis
hristov [Wed, 9 Nov 2011 14:51:25 +0000 (14:51 +0000)]
I have put as default the current parameterization of V0 and SPD --from run 166530 and 166532
Highly non-perfect but since the ranges are very different from last year better than the old default
richterm [Tue, 8 Nov 2011 10:10:29 +0000 (10:10 +0000)]
bugfix, follow up for #83123: two independent HLT systems with separate configuration handlers still had the problem that the configurations have been mixed because of using only the global configuration handler during extraction of sources; now the configuration handler is required as parameter in AliHLTConfiguration::ExtractSources and all global backdoors are removed
jthaeder [Tue, 8 Nov 2011 00:06:14 +0000 (00:06 +0000)]
Disentangled usage of HLT and offline one pad clusters
- added extra option the recoparam
Implemted options for rec.C for input switching of TPC reco
//--------------------------------------------------------------------
// Options for the TPC Reconstruction in rec.C
//
// 4 options can be set to change the input for TPC reconstruction
// which overwrites the usage of fUseHLTClusters of the AliTPCRecoParam
//
// 1) useRAW - use RAW, if not present -> do nothing
// 2) useRAWorHLT - use RAW, if not present -> use HLT clusters
// 3) useHLT - use HLT clusters, if not present -> do nothing
// 4) useHLTorRAW - use HLT clusters, if not present -> use RAW
//
// -> The current default is useHLTorRAW
//--------------------------------------------------------------------
richterm [Mon, 7 Nov 2011 20:45:48 +0000 (20:45 +0000)]
bugfix #83123: registration of configurations in multiple handlers. The bug caused configurations not to be registered in a particular configuration handler because the system tried to register the configuration in the global configuration handler. Furthermore, already existing configurations in the global handler interfered with the one going to be registered
richterm [Mon, 7 Nov 2011 20:32:33 +0000 (20:32 +0000)]
store all HLTOUT data blocks of an event in one DDL and use a round robin schedule to select the DDL; the old mode where blocks are distributed over all available links can be enabled by option '-distribute-blocks'; updated documentation
richterm [Mon, 7 Nov 2011 20:17:59 +0000 (20:17 +0000)]
automatically added data sink components are now added directly to the internal instance of configuration handler (required for bug #83123); cleanup of TObjString usage without functional change
prino [Mon, 7 Nov 2011 16:32:15 +0000 (16:32 +0000)]
SDD preprocessor updated to use new procedure to recover drift speed calibration for modules whose injectors were damaged during the beam loss event (Francesco, Gian Michele)