Software phone meeting on 10 October 2005

Present: Bruce, Dimitri, Florian, Gilles, Murrough, Norman, Richard, Steve.

Software for cable installation testing

Florian reported that the only recent change in the software was fixing the handling of full buffers in the RemFPGA. The test setup at Heidelberg is now using external L1As from TTCvi in a separate crate, as at CERN.

We discussed what is still needed when we move from VME to ROD readout. The code will need to handle proper ROD fragments. At low rate the ROD could handle larger than the standard 5 slices.

It will help the transition if the code was factored so that the analysis part expected our standard byte stream decoded objects and if the VME readout part produced them. Steve should send Florian some details and examples to follow.

At CERN we will need a reliable frozen release at all times for the tests. Development will happen in parallel and we need to define times when we can change versions. It may be that the ROD readout is only used for LAr tests.

It was proposed that we have another phone meeting to follow progress, perhaps on Monday 24 October.

Automated tests

At a previous meeting Murrough had mentioned some work on scripts for use in automated series of tests.There are TDAQ command line tools for sending commands to run controllers. Additional commands to set IS variables are now included in our iguiL1Calo package.

Following prompting by Gilles, some example scripts are now included in a new "automatedTests" package in CVS. Gilles has successfully adapted the example for use in CPM tests loading first a ramp and then physics data. Histograms produced by the tests still need manual intervention though.

It was mentioned that this scheme of changing IS variables is not really the ideal way - we should instead change the database.

ROD Crate DAQ

Murrough has been doing some further code development of new ROD crate DAQ run controllers - adding the VME readout. However this has yet to be tested.

Databases

Richard reported that he still was having no success in writing to a COOL/MySQL database at RAL. He was about to try the newer version of COOL installed with the new TDAQ release.

After the meeting it appears that he has finally succeeded in writing some data!

It was suggested that we need a dedicated meeting to discuss a wide range of details of our intended use of the database: from packaging of our software to organisation of the data. One possibility is the day before the next joint meeting at CERN, is Tuesday 8 November.

Preparations for next TDAQ release (tdaq-01-04-00)

Bruce, Steve and Murrough have tried using the new TDAQ release, after Steve made changes to handle the new version of the eformat library. However in tests on the morning of this meeting there were still some unresolved database problems and we did not succeed in getting the ROS going.

Two days later though, Bruce reports that he has fixed these problems and our software including ROS and checking via kicker now runs under the new release.

We discussed the priority (or otherwise) of making another TDAQ migration given other worries. The feeling was that the transition should not be hard, does not affect current PPM test code and that we might need to make it soon anyway if we wanted any chance of combined runs at CERN with TileCal who have already migrated. The consensus was that we should in fact try to make this change soon.

AOB

Apart from the proposed software phone meeting and dedicated database meeting mentioned above, we also discussed the format of the software session at the joint meeting. It was felt this should have more of a slant towards the future than status reports, and be less informal with some slides to guide the eyes.