DØ Level 3 Algorithms Meeting: 19th June 2002 in the Farside

Trigger list changes

Following several ad hoc meetings and discussions we came up with a set of proposals that were presented to the trigger board. Here are the talks:

L3 tracking test run results: Daniel Whiteson
L3 primary vertex tool perfomance: Chris Barnes
Strategy for EM triggering: Volker Buescher

A summary of the ensuing discussion can be found in the trigger board minutes.

Talks on CPS

Status of CPS: Drew Alton.
Studies of L3 CPS performance: Andre Turcot (waiting for transparencies).

Action items/questions for L3 code (Andre):

Code needs to be run on a large sample of recent data to check stability, timing and memory consumption.
Correct map for real data needed. Is this part of the release?

Other Discussion

Efforts to get l3fanalyze running on the offline farms:
Jon reported that thanks in part to Elizabeth Gallas' help he now has the complete chain to start with the run number, get the corresponing trigger list from the runs database, and then extract the list of tools and types.
Yann Coadou reported that the l3fanalyze roottuple for the special run taken for L3 tracking had a size of 10 MBytes/event! Most of the volume is taken up with the debug information produced by the SMT, CFT and CAL unpackers. (The run was taken with all events mark and passed so this information appears on every event.) We agreed that the sensible default behaviour should be that the examiners for the unpackers should NOT be run. Jon will change the relevant .rcp parameters to make it so.
Release issues:
- Martijn Mulders reported some strange differences in the muon segments/tracks seen in the l3fanalyze roottuples between p11.08 and p11.09. Suspicion is that the L3 reconstruction code is working OK, but that there are issues with the analyze behaviour that need to be understood.
- Yann Coadou reported problems with the parameter "threshold" not being defined in the latest version of tools.rcp. Note added later: Jon has fixed this. Change has gone into p11.08.01 and will get into p11.09 in the next build.
- Daniel's road change to allow the tracker to work efficiently when relatively high p_t cuts are specified did NOT get released. It WILL get into the next build of p11.09.
Conclusions for online running:
- p11.08.01 could be used online to test global_CalMuon-7.40, which has the purely technical changes to L3 parameters agreed at last week's trigger board.
- p11.09 will be needed for L3 tracking to work.

Scribe: Terry Wyatt.