OWG Meeting 24-Oct-2012

Agenda

Oktoberfest starts at 3pm!

• Announcements
  • Ed will (tentatively) talk more on FPGA programming in Oct/Nov
• Review of minutes from 12-Sep-2012 meeting
• Networks in Hall D - Hovanes, Elliott
• First effort on raw data tape simulation - Dave L, Elliott
  • multiple data formats
  • reconstruction in JANA
  • data challenge
  • translation table and tagger
• JInventory database - Elliott
  • conventions
• Counting House status - Hovanes
• DAQ group projects- Bryan, Dave A, Graham, Carl, Vardan, Elliott
  • CODA3, EVIO, ECS/CSS, disentangler, readout list, Java/cMsg GUI
• EPICS - Hovanes
• Controls - Elliott
  • Solenoid test: Review 6-Nov, PLC programming, PXI status, QD, connection to Cryo group PLC, etc.
• Electronics, trigger and DAQ status reports - Chris, Alex, Dave, Fernando, Ben, Ed, William, Bryan
• Mantis Task Tracker
• Midas Elog

• Calibration/Alignment issues?

Time/Location

1:30 PM Wed 24-Sep-2012 CC F326

Next Meeting

1:30 PM Wed 7-Nov-2012 CC F326

Minutes

Present: Elliott W, Beni Z, Bryan M, Hovanes E, Curtis M (remote), Dave L.

Networks in Hall D"

Hovanes reported that the network switch is installed and working in the Rack Room, and that we will meet with the Networking group this Fri. He proposed a number of subnets for Hall D (see his slides) that separates DAQ, control, test, and general functions. It is important to allocate enough addresses on each subnet since modifying the subnet definitions in the future is possible but somewhat of a pain. Note that this switch only deals with Ethernet traffic, we will purchase another system to handle Infiniband (used in the DAQ system).

Simulated Raw Data File

Elliott reported first results using Dave Abbott's "mc2coda" package that writes out detector hit information in the CODA online data format EVIO. See talk above for more details:

• mc2coda currently simulates block mode readout with block size n = 1, will implement n>1 eventually.
• Data format exactly as expected out of the front-end boards, as packaged by the event building system.
• Thus this format is verbose and contains redundant information.
• This is the format we'll use in the early debugging stages, but not for production running.
• Dave L will upgrade DANA to accept both blocked-mode and compressed data.
• Current version max 5 Hz, Dave is working to speed this up (required if we want to process 10**9 events!).
• Elliott wrote a compression algorithm that eliminates all redundant and extraneous information.
• Average event size for uncompressed format is 28 kBytes, compression drops this almost by a factor of three to around 10 kBytes.
• It is possible the compression algorithm is too strong.
• Dave L generated the input data: low-luminosity, smeared, L1 filter applied.
• Translation table code in rawevent plugin needs to be migrated to main DANA system.
• The first offline data challenge will NOT use simulated raw data, as it appeared too late, but future challenges will use it.

Counting House

The big UPS is powered but power is not enabled yet to the racks, not sure why. Counting house printer being installed in the Hall, request by Tom.

EPICS

Hovanes is developing a package that automatically generates IOC databases from Allen-Bradley XML program dumps. The A-B XML code is convoluted and not easy to understand, but he hopes to have it done in the next week or so.

Solenoid Test

Cryo review 6-Nov, pump and purge scheduled for mid-Nov, cooldown for late Nov...everyone is VERY busy!

At this point we broke to go to the Oktoberfest.