Title :
Performance of the L3 second level trigger implemented for the LEP II with the SGS Thomson C104 packet switch
Author :
Blaising, J.J. ; Cai, X. ; Flour, F. Chollet-Le ; Cruz, J.C. ; Daguin, G. ; Degre, A. ; Martin, B. ; Masserot, A. ; Perrot, G. ; Zhu, M.
Author_Institution :
IN2P3, LAPP, Annecy-le-Vieux, France
fDate :
8/1/1998 12:00:00 AM
Abstract :
The L3 experiment is one of the four experiments collecting data at LEP. For the LEP phase II, the second level trigger has been upgraded to a network of 28 ST T9000 transputers and 2 ST C104 asynchronous packet switches interconnected by IEEE1355 links. It collects trigger data at each LEP crossing (22 μs), builds-up the trigger data block, processes it and rejects online the background events in a few milliseconds. The L3 data acquisition has been running with this system since July 1995. In the data-taking environment and for different hardware and software implementations, the event building throughput rate has been measured. A bandwidth of 5.9 Mbytes per second per link has been measured in a configuration with 1.2 sources and one processing unit connected with 2 links. The expected global throughput of 70 Mbytes per second has been measured in a farm of 6 processing units. While varying the number of sources and destinations, we didn´t observe any significant bandwidth loss. Nevertheless performance relies strongly on some software implementation choices, which are presented and discussed
Keywords :
colliding beam accelerators; data acquisition; electron accelerators; high energy physics instrumentation computing; nuclear electronics; storage rings; switches; transputers; trigger circuits; IEEE1355 links; L3 second level trigger performance; LEP II; SGS Thomson C104 packet switch; T9000 transputers; asynchronous packet switches; background events; data acquisition; electron storage ring; event building throughput rate; global throughput; software implementation choices; Bandwidth; Data acquisition; Detectors; Event detection; Hardware; Packet switching; Software maintenance; Software measurement; Switches; Throughput;
Journal_Title :
Nuclear Science, IEEE Transactions on