CERN QMUL
Home
|
L1Calo Software meeting on 5 March 2003
Present: Bruce, Cano, Eric, Gilles, Murrough, Norman, Sam, Steve, Thomas.CMM software status: Norman
The CMM register model has been updated - though Ian has subsequently sent Norman further changes. The CMM module services were updated as part of recent changes to the module services API. Some methods still need to be populated with code. The first readout checks will be done via HDMC, after that its still undecided.CPM software status: Gilles
Extra functions (eg setting masks) have been added to the CPM module services. There is still more information from the database to be incorporated, eg loading firmware. The test programs are now in a separate cpmTests package and are running comfortably via run control but still with a private version of ttcviServices. It is hoped to test the TTCrx I2C interface soon. Another run control and database integration session would be useful.JEM software status: Cano
Standalone software for input synchronisation scans now exists and uses two LVDS DSS daughtercards. Cano has also implemented scans using large numbers of test vectors via scripts. In future this should be done via the run control. Also some reorganisation of code out of the Jem class into test programs is needed (for tests with more than one JEM).Tests reading the outputs from the readout controller have also been done, but only with visual comparison. There is also software to test the fanin/out via the backplane (but so far only to the same JEM via loopback cards).
Recent changes to the database to describe module firmware has been tried - and bugs discovered in the database code (though an older version was being used). The database also needs to allow the results of the input synchronisation to be stored as calibration data which could the be loaded - after some more work on JEM module services.
Storing firmware
We had a short discussion on how we should store our firmware binaries. At the moment Cano keeps them in our HDMC CVS repository area. Firmware binaries for other modules are not (yet) kept in any single central place.
A number of suggestions were raised in the discussion. We could store everything on our CERN AFS space, or in CVS somewhere. Although CVS is not optimised for keeping binaries, if we also put the firmware under CMT we could neatly combine software and firmware releases.
The CVS options include the HDMC area (but this will hopefully be merged into the l1calo area someday). Alternatively we have a separate firmware CVS tree, so far used only for firmware development sources. Or we could start using the TDAQ CVS server at CERN for this - though this is probably intended only for software source code which takes up much less space. Lastly, the RAL engineers are investigating CVS-like solutions for storing versions of their development source. We need to find out what the status of this possibility is. It may provide a nice solution for the binaries too?
We didnt reach a decision, but should keep the issue in mind.
DSS, ROD, ROS: Bruce
Following extensive discussions at previous meetings, the module services (DaqInterface) API has been extended and Bruce undertook the task of updating all module services packages according. The new DaqInterface is fully Doxygenated.As previously reported, Bruce restructured the DSS module services to support two daughter cards of the same type. This is now being used successfully. A problem with Glink source cards has been resolved.
Testing of new ROD firmware variants continues. There is a new person at RAL to help with this, so it should go faster in future. Testing variants would also be helped if the firmware could be reloaded via VME. Dave Sankey will look into this.
Bruce listed some of the "to do" tasks. We hope to create an "allModules" package to provide an HDMC version with all our modules. The "kicker" program(s) should also use the monitoring framework to present events for comparison. If Dave Sankey takes over some of the ROD testing, Bruce would hope to look at L1A generation via the DSS.
Our readout and "monitoring" (or more properly data acquisition) requirements will be discussed with Dave Francis et al at CERN in the next few weeks. We will (hopefully) want to read out multiple RODs soon, but we shouldnt demand this until we really need it.
Database and run control: Murrough
Murrough gave a short talk summarising the recent extensive changes to the database code and also showing how the run control uses the old and new methods of the DaqInterface. Some tuning of this is probably desirable. It was also suggested that the new run type options in the database should also allow module firmware to be reloaded when the run type is changed.Simulation: Steve/Sam
Steve has been making continued udpates and corrections to the CPM simulation. He has unfortunately had no time to look at the JEM simulation code received from Daniel and Asa.Sam reported that the JEM simulation now compiles and runs, but it is still missing some of the algorithmic content in the process elements. The aim is to get the software to the point where it takes input and gives correct real time output. Then responsibility will be passed to Steve to integrate it into the CMT framework. The timescale for this is the end of March.
Sam should get hold of the energy summation algorithm from Juergen. The missing Et lookup tables need to be calculated but may be taken from the database eventually.
Actions from previous meetings
Norman has obtained NT accounts for the Mainz/Stockholm visitors. He needs to check whether access through the firewall via accounts on the RAL bastion server are included.OS/Online migration
There are two new PC servers at RAL. Bruce proposes to load one with RedHat 7.3 and later versions of other software so we can try to migrate our standard environment to more recent versions.Next meeting(s)
To be arranged. Maybe just during the next JEM visit to RAL?Actions
Actions from this and previous meetings:- Bruce/Murrough: write and circulate proposal for handling multistep runs in DaqInterface and run controller
- Bruce: complete Module Service document and submodule syntax description
- Bruce/Murrough: define required HDMC improvements
- Murrough: tackle project management issues: eg schedule (Gantt chart, stages of evolutionary delivery), risks, etc
- Murrough: update database user guide
- Murrough: complete run control design document
Last updated on 08-Mar-2003 by Murrough Landon