OWG meeting 18-Apr-2008

From GlueXWiki
Jump to: navigation, search

Agenda

  • Announcements
  • Review of minutes from 11-Apr-2008 meeting
  • CODA3 components: EMU - Graham
  • More on event size study - Alex S.
  • Status of writing assignments and other preparations for subsystem review - many people
  • Electronics - Status of documentation - Fernando


Time/Location

Note special time - 10am Fri 18-Apr-2008 CC F326

Due to scheduling problems, this will be a joint meeting with the Electronics group.


Announcements

Subsystem review will be Wed,Thurs 7,8-May. GlueX collaboration meeting scheduled for the end of that week.

Information concerning the charge, agenda, dry run schedule, etc. can be found on the May 2008 Subsystem Review page.


Elke's Deadlines

  • Final presentations to Elke: 5-May close-of-business
  • Pre-brief material to reviewers: 30-April
  • Pre-brief material to Hall-D group to read: 23-April
  • Pre-brief outlines: 11-April


Next Meeting

May not happen due to dry runs


New Action Items from this Meeting

Minutes

Attendees: Elliott W, Ben R, Dave A, Simon T, Beni Z, Dave D, Time W, Chris C, Eugene C, Mark I, David L, Graham H, Ed J, Carl T, Vardan G, Elke A, Elton S, Hai D, Alex S, Jim S, Fernando B.


May Subsystem Review


CODA3 and EMU Design

Graham outlined the CODA 3 architecture and the discussed role of the Event Management Unit (EMU).

  • CODA3 = software toolkit plus hardware recommendations.
  • Graham presented history of versions 1 and 2 of CODA, starting in the early 1990's.
  • Version 3 for upgrade, includes lessons learned over past decade.
  • V2 max 32 crates, single event builder, VxWorks only in ROCs, one event recorder, no pipleline readout, etc.
  • V3 no crate limit, multiple/parallel event builders, Linux in ROCs, multiple event recorder, pipelined/blocked readout, etc.
  • Do no want to lock into any particular technology for CODA3, use framework + plugin strategy.
  • EMU (Event Management Unit) framework for event building/manipulation.
  • Uses XML extensively for configuration (input, output, which plugins to use, etc).
  • Plug-ins: read, send, store, build, L3.
  • Original EMU in C++, Java version in progress.
  • We discussed a number of event building architectures and crossbar strategies.
  • Can load-balance via crossbar configuration.
  • EMU flexible, adaptable to multi-core era.
  • EMU prototypes being tested, no real event building yet, but architecture working.
  • Full test on LQCD cluster in Sep.
  • We had a discussion on blocked readout in front-ends and where this should be disentangled.


Update on Event Rate/Size Study

Alex gave an update from his report last week. See link above for many details.

  • Considered signal plus EM and hadronic backgrounds.
  • EM using 100 ns gate, 1.1 MeV photon cutoff.
  • Used simple cuts. Original CNU cuts tuned w/o EM background so not very effective, need updating.
  • EM 78% have no ST signal, hadr 98% have ST signal.
  • 93% overall hadr efficience w/Alex's cuts.
  • 82 kHz EM plus 98 kHz hadr = 180 kHz total L1 trigger rate with Alex's cuts.
  • This work adequate for May review.


Electronics Update

  • A lot of background material for pre-brief can be found on the Electronics page.
  • Information on grounding in preparation.
  • Overview document coming.