New GSI Analysis System GO4  

Participants: W. Brüchle, H. Essel, H. Göringer, R. Holzmann, P. Koczon, I. Koenig, W. Koenig, C. Kozhuharov, N. Kurz, W.F.J. Müller, M. Richter, E. Stiel

Distribution: R. Barth, H. Brand, P. Braun-Munzinger, S. Glückert, N. Herrmann, S. Hofmann, M. Kaspar, H. J. Kluge, U. Krause, U. Lynen, V. Metag, G. Münzenberg, V. Schaa, C. Scheidenberger, C. Schlegel, K. Schmidt, K.H. Schmidt, H. J. Specht, R. Steiner, K. Sümmerer, H. Winter, Juha Äyastö (Jyväskylä), G. Bollen (ISOLDE), R. Fox (MSU), E. Grosse (FZR), Th. Nilsson (ISOLDE), M. Schlett (FZR), H. Simon (ISOLDE), K. Suzuki (Niigata), Ken Teh (ANL), Piet Van Duppen (Louvain-la-Neuve)

H. Essel, H. Göringer and P. Malzacher (all GSI) had been in Chicago on the CHEP 98 meeting. Following individuals showed up interest in GO4 after talking with Hans Essel:

• Ron Fox (MSU)
• Ken Teh (ANL Physics Division) is very much interested in GO4. He even would like to spend a year as a guest at GSI working on GO4.
• Rene Brun (CERN IT/DI), father of ROOT. He is currently at Fermilab working together with experimental groups in the decision of taking ROOT as a standard. Rene told also that ROOT has most of its code written for use with threads, although the current version does not yet show any hooks.

The CORBA standard seems to become a general tool. In version 3.0 there should be an interpreter (CorbaScript) for fast prototyping.

All details of the HADES experiment, the detector set-up, the data acquisition, the trigger system, and the analysis can be found under: http://hades.gsi.de/ and http://www.e12.physik.tu-muenchen.de/hades/.

The High Acceptance Di Electron Spectrometer (HADES) is a detector system for lepton pair spectroscopy presently built up at GSI. As leptons do not undergo strong interactions but interact only electromagnetically they are considered penetrating probes of compressed hadronic matter formed within the collision zone of relativistic heavy ion reactions. The very weak final state interaction allows a measurement of vector meson masses in the hot and compressed hadronic environment by reconstruction of the invariant mass from the lepton momenta. This provides the possibility to test theoretical predictions for a partial restoration of chiral symmetry at elevated temperatures and baryon densities.

HADES consists of various detector systems: RICH (Ring Imaging Cherenkov Detector), MDC (Multi wire Drift Chambers), META (Multiplicity/Electron Trigger Array), TOF (Time Of Flight Wall) all together with a superconducting torus magnet.

The main experimental problem will be the reduction of a high background. Therefore a three level trigger system will reduce the primary data rate of 10**6 reactions/spill (1 spill every 3 s) to 10**2 events/spill.

• The 1st level trigger realized by a multiplicity box (hardware) reduces the event rate by a factor of 10.
• On the 2nd level trigger FPGA logic will reduce the event rate by another factor of 100.
• The 3rd level trigger to be realized with software running on specific DSP boards will give a further reduction by a factor of 10. The 3rd level trigger is not yet implemented. It might be done by off-line analysis in the first year.

The final event rate of 100 events/spill has to be written on tape. The calculations of an average event size after data compression (zero suppression) leads to 900 kByte/spill data rate on tape. With this assumption the raw data volume produced by HADES in one year will be about 1 TByte.

The analysis structure follows the layout of the detector components:

• cells within each module. The raw data coming in each cell will be calibrated.
• modules within each sector. Several cells within a module define valid hits. Hits define track pieces and finally tracks.
• 6 sectors. Sectors are physically isolated and can be analyzed independently.

The various parameters of each level of analysis (like calibration factors) are stored in an Oracle data base. During the analysis they will be stored in a memory-mapped data base. Each level of analysis will allow writing and reading of DST (Data Summary Tapes).

The collaboration has opted for ROOT as its standard analysis software package. All analysis classes are derived from the ROOT base class TObject, such that the ROOT dictionary RTTI offers:

• automatic generation of streamer I/O functions
• analysis via ROOT Trees
• HTML documentation
• on-line class description
• context menus, browser, etc.

For on-line analysis a Tree Data Structure has been designed to follow the detector geometry and the various analysis levels.

First experiences with ROOT are positive:

• reasonable performance can be achieved
• flexible data structures allow on-line analysis (ntuples)

The C++ command language is new to most users and requires some effort to get used to. The development of a GUI could greatly ease the transition to a ROOT-based analysis.

M. Richter, October 1, 1998

Downloads-Manuals-References

GSI Helmholtzzentrum für Schwerionenforschung, GSI
Planckstr. 1, 64291 Darmstadt, Germany 
For all questions and ideas contact: J.Adamczewski@gsi.de or S.Linev@gsi.de
Last update: 27-11-13.