TrigExPartialEB: Add more MTCalibPeb options for performance studies
With this change, when running MTCalibPeb options for performance studies, we will be able to:
- create any given number of chains in an automatic way (e.g. 1000)
- specify hypo tool options from command line, e.g.
-c 'RandomAcceptRate=0.5;BurnTimePerCycleMillisec=50;'
- use CPU cruncher instead of sleep to better simulate load
Tested in several ways, including 10 concurrent chains with 50% accept rate per chain and crunching enabled, running with 2 forks, 6 threads each, 2 events on the fly:
athenaHLT.py -l VERBOSE -c 'concurrent=True;num_chains=10;RandomAcceptRate=0.5;Crunch=True;' --imf --threads=6 --concurrent-events=2 --nprocs=2 --number-of-events=50 --file=/cvmfs/atlas-nightlies.cern.ch/repo/data/data-art/TrigP1Test/data18_13TeV.00360026.physics_EnhancedBias.merge.RAW._lb0151._SFO-1._0001.1 TrigExPartialEB/MTCalibPeb.py