1 / 9

Hbd Software Update

Hbd Software Update. By Sky Rolnick 8/04/09 Hbd Group Meeting. Issues with Simulation (solved ✔ ). Detector Geometry Before we were seeing inconstancies between simulation and data relating to differences in coordinate systems. This issue has been solved. ✔ Missing Track projections

donkor
Télécharger la présentation

Hbd Software Update

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Hbd Software Update By Sky Rolnick 8/04/09 Hbd Group Meeting

  2. Issues with Simulation (solved ✔) • Detector Geometry • Before we were seeing inconstancies between simulation and data relating to differences in coordinate systems. This issue has been solved.✔ • Missing Track projections • It seemed that all track projections in cgl were -9999. This was due to missing HbdMiniCellList node in simulation. Cgl now checks for either HbdMiniCellList or HbdCellList. This issue has been solved. ✔ • Missing Hits in West Arm • HbdFinalSimulator now uses hbdDetectorGeo for its coordinate transfomations. This resulted in a lack of hits in the West arm. This was traced down to different definitions of side. In HbdFinalSimSupport “side” means N/S, but in hbdDetectorGeo “side” means E/W. This issue has been solved. ✔ Sky Rolnick UC Riverside HBD Group Meeting

  3. Issues with Detector Geometry (solved ✔) • We now have consistent geometry definitions in both simulation and data. • All code should now use hbdDetectorGeo as HbdFinalSimSupport is considered outdated. • All functionality of SimSupport has been ported over to DetectorGeo. Many new methods, ie. getPadInfo, getPadLocal, getPadGlobal, PanelToSect, FindNearestPanel, FindNearestPad, DistanceFromPanel, etc. • Just a warning, some of the functions have changed. • Example void GlobToLoc(double GlobX, double GlobY, double GlobZ, double& LocY, double& LocZ, int ipanel); (old) void GlobToLoc(double GlobX, double GlobY, double GlobZ,                           double& LocY, double& LocZ, int& LocSect); (new) Sky Rolnick UC Riverside HBD Group Meeting

  4. Issues with HbdHnSClusterizer(solved ✔) • Small changes to HnS Clusterizer… • getThreeTupleHub now compares next cell to WORST of three, not BEST of three. ( this was the right thing to do, but you guys checked in a bug.) (solved✔) • getThreeTupleSpoke had a bug, since it returned a pointer, the contents of object would sometimes change leading to misidentified spoke. This bug was fixed many months ago, but somehow resurfaced. (solved✔) • COG (Center of Gravity) had some bugs. Besides not being zeroed correctly, it had issues when clusters were defined near boundary. (solved✔) Sky Rolnick UC Riverside HBD Group Meeting

  5. New HbdCluster Object? • We discussed last week the addition of HbdCluster object to the hbd package. • There is some concern that this would create some redundancy with HbdBlob object. • HbdCluster would essentially be a HbdBlob with the addition of a list of HbdCells and some additional methods like sort(), clone(), etc. • Question is whether to introduce this as new base class or simply extend HbdBlob base class? • I envision HbdCluster to be used internally for clustering but HbdBlob to be written out to memory or disk. Does that sound reasonable? Sky Rolnick UC Riverside HBD Group Meeting

  6. Remaining Issues with HbdFinalSimulator • This code still crashes when trying to process central AuAu events. This is the infamous “maxtracks exceeded” error mentioned many months back. • There are several things about the code I don’t understand. For example why to we read in variables from dio_ptrkorigin and then immediately override them with variables from dio_ptrkstack? • Also seems that HbdPISAHit is defined as a pointer to memory, but then being accessed as an array with size PISAnHits?? Seems very dangerous to me. • Many other issues remain with this code, but I’m not an expert, and am not ready to rewrite all this yet. Some input? Sky Rolnick UC Riverside HBD Group Meeting

  7. Remaining Issues with Hub n Spoke • I have defined hub and spoke as global pointers. This is very dangerous and bad coding. I will change this to be passed by reference in the near future. • It may be possible to up-convert from v22 to v24 without the use of v24maker. This seems to be what is done in several Recalibrators. Not sure how this works? i.e. baseclasses.insert("PHCentralTrackV24"); • Still need to work out how to deal with overlapping clusters. Right now if two tracks are very close together, we will get two cluster with overlapping cells. • It is also possible to have multiple tracks associated to a single cluster, ie. for Dalitz or conversions within Hbd. Right now this will be identified as two clusters. • Still need to implement analysis code which can provide histograms with basic track matching, electron pairs, and single & double efficiency. Sky Rolnick UC Riverside HBD Group Meeting

  8. Plans for next week …. • Fix any remaining bugs with HbdFinalSimulator so that central AuAu events can be processed correctly. • Reprocess all monte carlo data through new HbdFinalSimulator since this has new geometry updates. • Work on matching single electrons to Hbd clusters. This will probably require the use of PairObj to identify electrons from J/ψ and treat them as single electrons. Try to get out an single electron efficiency. • Begin to study clustering algorithms in higher multiplicity using either generated Monte Carlo or EventAccumulator with run9 data. • Begin looking for dead strips using pad frequency normalized to pad area. Give us an idea of where the dead strips are located and how much this affects our efficiency. • Eventually study Hadron rejection as a function of gain using several different modules at various gains for different runs. Is there enough 500 GeV (+-) data for this? Sky Rolnick UC Riverside HBD Group Meeting

  9. The End …. Sky Rolnick UC Riverside HBD Group Meeting

More Related