Software phone meeting on 14 September 2006

Present: Bruce, DaveC, DaveS, Damien, Florian, Gilles, Marianne, Murrough, Norman, Rainer, Steve, Victor.

Status reports

Birmingham

Gilles has committed the CANopen software developed by Chris to CVS under the firmware directory.

Heidelberg

Florian is now using GNAM. At least the DAC scan and pedestal runs should be ready for tests next week. Other tests would need signals from TileCal. He has a number of GNAM issues - see below.

Victor has done some work on converting his monitoring to GNAM but this is not yet finished. He been working on the channel mappings document and will return to work on GNAM after he returns in October.

Mainz

Rainer asked if Mainz could have a DSS GIO card to set up a system like the one at Birmingham to drive the trigger. Bruce said there were spares at RAL.

RAL

Dave has written a little tool to take test vectors generated by the online and run them through the VHDL simulation.

CERN

Marianne raised the issue of how we will deal with lots of histograms in the final system: her present work creates many histograms per JEM already. The conclusion from a little discussion was that we should expect the monitoring data archive to cope.

Bruce has prepared the updates to modules services for tdaq-01-06-01. He will tag all packags and made the updates to all services packages in the next few days.

Murrough will try tdaq-01-06-01 in USA15 (so far only tried the previous tdaq-01-06-00) and should then install it at RAL and make a tarball for others to use.

Steve has implemented the PPM compressed readout format in the simulation and the bytestream decoder. The new code is fairly well tested. He has also started using the new channel mappings in the bytestream decoder package so the genuine eta and phi is now available (except in the forward jet region).

Following this, Bruce has added some suitable test vector generator patterns to provide realistic data for the PPM compression algorithm.

GNAM issues

Florian had a number of GNAM related questions and problems which included:

• how the VME decoder was going to be made available
• how to specify where to store results
• his GNAM library crashes on a second configure step
• no combination of prepublish and endOfRun seems to allow histograms to be published after the stop transition
• the complete set of GNAM entry points should be added to the HistogrammerBase class.

We agreed we would have a GNAM experts week in the week beginning 23 October.

Disk space at point 1

We ahve been asked to reduce our use of the home directory area at point 1 - this is limited. Instead we should store all our software releases, both official and private ones, on the /det/l1calo area. Results data should be stored in the database...

Next meeting

The next phone meeting will be on Thursday 28 September at 11:00 CET (10:00 GMT).