Investigations of Parameter Space

On this page I've collected some first results from probing the parameter space of ATRECON. All these plots are highly preliminary.

	Dependence of xKalman time from number of seeds for 30GeV pT electrons and dijets at high luminosity. By using the ROI information from LVL2 or by increasing the threshold for seed clusters in the ECAL, the number of seeds can be reduced. For an ECAL threshold of 15GeV the numbers of seeds is halved compared to 5GeV, reducing the maximal reconstruction times considerably, without loosing anything in efficiency or rejection.
	The xKalman default size for the track search around an ECAL seed is 0.5x0.5. By lowering the size of the road to 0.1x0.1, a significant gain in reconstruction speed is found for high luminosity. As expected, the gain at low luminosity is smaller. The upper plot shows the timing of xKalman for electrons, the lower plot that for dijets. The bars show the value in which 50%/80%/95% of LVL2 accepted events can be reconstructed. The reconstruction efficiency and jet rejection is not affected by the smaller seed.
	A study of the minimum pT for reconstruction of tracks with xKalman was done for dijets with pT>35GeV. This plot shows the dependence of the reconstruction time needed vs. the ptmin set. The bars show the value in which 50%/80%/95% of all events can be reconstructed. No studies have been done concerning the efficiency/rejection issue. However the comparison with Pascal's work was done with minpt=5GeV. Therefore I don't expect a large difference in efficiency/rejection for lower ptmin settings.

More information can be found in the DAQ-note ATL-DAQ-2000-007.