CERN QMUL
L1Calo Software video conference at Bham/HD on 4 February 2002
Present: Bruce, Gilles, Karsten, Murrough, Oliver, Steve.Actions from previous meetings
- Some actions are addressed in the progress reports below.
- Other actions are carried over...
Murrough: run control, databases and CMT
Murrough briefly summarised work on the run control reported at the previous (UK based meeting). Since then he has done more work on adding IGUI panels to set/display run parameters stored in IS variables. More work on the database remains a high priority.He then presented a talk on using CMT. He has copied the working model from the Online group, some scripts from the DataCollection group and has now written CMT "requirements" files for each of his packages. It is proposed that our other recent and forthcoming packages should also be handled by CMT. HDMC, requiring generation of Qt "moc" files is more complex and can perhaps be left until later - though maybe someone has already developed suitable CMT code for Qt?
Bruce: module services, ROS, system management
Bruce has been working on documentation describing the submodule syntax he has added to HDMC and, with slightly lower priority, the module services package. He promised to finish at least the first by the end of the week.The promised beta of the official distribution of the ROS software has not yet been released. Bruce still intends to test it as soon as it is available.
System management: Bruce has had to install NT on atlun01 at RAL in order to run the Chipscope diagnostic package. He hopes to regularise all the Linux systems in the lab, probably moving to RedHat 7.2. (During the discussion it was reported that Birmingham might stay at RedHat 6.2, but after the meeting it emerged that the local Birmingham system manager was happy to install 7.2 on the Birmingham machines).
Lastly, Bruce reported on the continuing ROD tests. Tests with Chipscope and a logic analyser are proving useful in pointing out remaining firmware problems to the engineers.
Gilles: CPM module services
Gilles has continued to develop the module services for the CPM. He has not yet stored them in CVS - it will probably be best to put module services developments in a separate package from HDMC.He has also obtained the CpChip code from RAL and has been developing a VHDL test bench so that they can run the CpChip simulation. This should be useful when testing the CPM (which is now being assembled).
VMM
In Normans absence, Bruce reported that Norman has been working to (retrospectively?) define the specification of the VME mount module (VMM) which we need to host 6U CPUs in our 9U crates.Steve: simulation
Steve has now added copious Doxygen style comments to his code and has produced extensive reference documentation. He has not quite finished an introductory document but expects to complete it soon. But there is already a web page describing the simulation package. He thinks the code for the simulation framework is now pretty stable. The most recent addition was adding support for changing trigger menus to allow scanning thresholds with constant input data.He has had one interaction from the Stockholm project students and has sent them more help. But we dont know how they are progressing with the JEM simulation.
Oliver: progress at Heidelberg
Oliver reported on recent developments at Heidelberg. The ASIC has now been submitted and is expected back in March. Meanwhile the MCM has been altered again (to use the correct eta,phi BC multiplexing) and is also in production.Heidelberg continues to suffer from shortage of people: Werner is leaving after completing the MCM. Dominique has presented his work on the RemFPGA and will also be leaving soon. His present design uses 40% of the Xilinx XCV1000E resources. All the devices for the final system have now been bought.
The priorities at Heidelberg are dominated by the Hardware programme. Oliver has to work on the firmware for a test board to read out the ASIC. The expected work on the ROD firmware to receive the RemFPGA output will delayed until the autumn.
Another urgent matter is interfacing the VME interface on the "Heidelberg home brew" CPU into HDMC. Related to this (but lower priority) we should consider moving to the newly proposed CERN API for VME access. Oliver would like to see these proposals.
Karsten: video memory
Karsten reported that he has completed his program to load the video memories. They can now load and playback all the measured calorimeter pulses.Status at Mainz
Thomas was unable to attend, but Murrough reported that they have the first JEM back at Mainz (a second is being manufactured) and that Thomas was using low level VMElinux driver calls to test its VME interface.ROD crate DAQ
We briefly discussed the ROD crate DAQ task force. Norman has been diligently reporting summaries of the task force meetings. One issue which arose there was whether HDMC could be included as a part part of the ROD crate DAQ suite. Norman had proposed a response offering HDMC "as is" but with the caveat concerning the level of support. Oliver said that he agreed with Normans response.Schedule
No progress: Murrough should chase Norman to return the draft project in MS Project format, together with hardware milestones.Next meeting
A few people will be at ATLAS week. Apart from that the next meeting will be the joint meeting in Heidelberg.Test Vector Brainstorming
After the video session, those in the UK held a brainstorming session to discuss the issues around handling test vectors. The results of this will be summarised and circulated later.Actions
Actions from this and previous meetings:- Norman: find/define remaining data formats
- Steve: complete introductory documentation for simulation package
- Bruce/all: investigate software development tools
- Bruce: complete Module Service document and submodule syntax description
- Bruce/Murrough: define required HDMC improvements
- Murrough: complete run control design document
- Murrough: write up test vector discussion as a document
Last updated on 11-Feb-2002 by Murrough Landon