DocumentCode :
2622524
Title :
The ALICE Dimuon Spectrometer High Level Trigger
Author :
Becker, Bruce ; Chattopadhyay, Sukalyan ; Cicalò, Corrado ; Das, Indranil ; de Vaux, Gareth ; Fearick, Roger ; Lindenstruth, Volker ; Marras, Davide ; Sanyal, Abhijit ; Siddhanta, Sabyasachi ; Staley, Florent ; Steinbeck, Timm ; Szostak, Artur ; Usai, Gia
Author_Institution :
UCT-CERN Research Centre, University of Cape Town, Rondebosch, 7700, South Africa
fYear :
2008
fDate :
19-25 Oct. 2008
Firstpage :
1846
Lastpage :
1850
Abstract :
The ALICE Dimuon Spectrometer High Level Trigger (dHLT) is an on-line processing stage whose primary function is to select interesting events that contain distinct physics signals from heavy resonance decays such as J/ψ and ϒ particles, amidst unwanted background events. It forms part of the High Level Trigger of the ALICE experiment, whose goal is to reduce the large data rate of about 25 GB/s from the ALICE detectors by an order of magnitude, without loosing interesting physics events. The dHLT has been implemented as a software trigger within a high performance and fault tolerant data transportation framework, which is run on a large cluster of commodity compute nodes. To reach the required processing speeds, the system is built as a concurrent system with a hierarchy of processing steps. The main algorithms perform partial event reconstruction, starting with hit reconstruction on the level of the raw data received from the spectrometer. Then a tracking algorithm finds track candidates from the reconstructed hit points. Physical parameters such as momentum are extracted from the track candidates and finally a dHLT decision is made to readout the event based on certain trigger criteria. Various simulations and commissioning tests have shown that the dHLT can expect a background rejection factor of at least 5 compared to hardware triggering alone, with little impact on the signal detection efficiency.
Keywords :
Clustering algorithms; Detectors; Event detection; Fault tolerance; Physics; Resonance; Signal processing; Software performance; Spectroscopy; Transportation;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Nuclear Science Symposium Conference Record, 2008. NSS '08. IEEE
Conference_Location :
Dresden, Germany
ISSN :
1095-7863
Print_ISBN :
978-1-4244-2714-7
Electronic_ISBN :
1095-7863
Type :
conf
DOI :
10.1109/NSSMIC.2008.4774751
Filename :
4774751
Link To Document :
بازگشت