Agenda:

  1. Software Review response                        -- Maurik, all
  2. Workshop
  3. Framework
  4. Software Schedule updates                       -- Maurik, all
  5. Project updates?                                      -- All
  6. Questions?                                              -- All

Please let me know if you would like to present something.

Meeting Notes:

blue highlighted items are new tasks/action items.

Responses-1

Prioritizing Tasks:  

A column is added to the software schedule showing the priorities of tasks. Most priorities are at "normal", the default. We need to pay attention to which tasks should have elevated ("high", "very high"=critical) priorities or low priority. 

Priorities will be set in two ways: From feedback from the detector groups and developers, and by looking at the timeline and the times that certain tasks need to be completed.

Data Quality Checks: 

Matt thought this was a great suggestion from the reviewers.
Sho told us that some of these checks are in place, but more are needed. 
Matt suggested that we need a Data Quality Manager, who implements and oversees the data quality. This person will need to interface with the Data Manager (formerly Homer, looking for new person) and the people from the detector subgroups. This is a very important job, but can be a lot of work. It would be good to have some level of automation.

Responses-2

Clustering Algorithms: Implementing the Jlab clustering algorithm is well on its way (Holly, Kyle). The algorithm will be implemented in Java shortly. There will also be a standalone version for comparison. A further step is to test the algorithms and to study the detailed behavior of the ECAL. 

Clustering in trigger: This is a different algorithm, because it needs to correspond to what runs on the FPGA in great detail. Kyle is reviewing the initial implementation from Sho and will make changes if needed. Kyle will also look at the trigger algorithm and make sure the diagnostics from the trigger are replicated in the software.

Globus Toolkit: Sho has already started looking into getting the Globus toolkit running at SLAC for improved data transfer speeds.

Responses-3

EVIO to Offline data mapping: Sho tells us this is in the code. It still needs to be interfaced with the Conditions Database.

Trigger monitoring: How to get the trigger monitor output accessible for comparison to our algorithms. This will be a sub-task of the task to check and update the implementation of the trigger in the offline code. (assigned to Kyle).

Interface with Slow Control DB: New task: write scripts to copy relevant information from the Slow controls DB to the Conditions Database, determine what data needs to be copied, resolve which run number is valid for the data.

Response-4

Leveling Manpower: This will be an ongoing process. As new members get familiar with the software environment they will likely be able to pick up more tasks. John mentioned that it looks quite scary, since some of the people, especially Pelle, Sho and Omar, are heavily involved with software but also have critical hardware tasks. If there is no hardware there is no need for the software, so hardware will take priority. Maurik will continue to actively monitor this situation and have as many tasks spread out over more people as possible. Specifically, it would be very useful to have an additional person work closely with one of the (overloaded) experts. This is already looking positive with Pelle working with Gabriel and Alessandra on the tracker alignment. Continue to monitor the software load for individuals and continue to solicit feedback from them. Also continue to recruit new contributors to take over tasks.

Response-5

Additional Milestones:  We will use "end of task" as additional milestones. Add other milestones where appropriate.

Response-6

Additional Liaisons:  

For DAQ - SVT, Pelle is already functioning in this role.
For DAQ - Ecal & Trigger, Kyle is currently already talking with the Jlab people, we'll ask him to take this job.
For Data Handling, this is what the Data Handling person is for.

We should be aware that there will be a second readiness review for HPS, conducted by Volker, in September. At that stage we want all our software to be in the deployment stage.

 

Framework

Maurik showed 3 slides outlining the Framework for HPS software, see slides with the agenda. All agreed that this was reasonable and represents our consensus of how we approach the software.

Schedule

Schedule updates will be pushed to the online schedule.

ECAL Digitization: Gabriel is making good progress, expects task to be complete in one week. 80% done.
DST:  Omar updated the hps_root dst maker, it now includes ECAL hits, he also reorganized the code a bit. Some further will be done. A follow up email shows where the information page is for this project: https://confluence.slac.stanford.edu/display/hpsg/HPS+Data+Summary+Tapes

 



  • No labels