- Table of contents
- 01/14/2013 Notes
Andrew, Jaroslav, Duyang, Zukai, Gavin, Xuebing, Yuri, Brian, Jan, Evan and Martin (taking notes)
DAQ Application Manager¶
- Yuri successfully implemented the DDT executable in the DAQ Application Manager (see Running the Trigger using the DAQApplicationManager).
- Andrew suggests that we should use this to write data.
- Yuri and I will work on getting this tested on the detector. We would like to run for a day without writing data to see whether the DDT interferes with anything that we are unaware of. Andrew would also like us to do this while it is writing data.
- Yuri and I will discuss a best time with Mat (we talked after the meeting and I will run this test tomorrow).
- Zukai would like to run his Supernova trigger for a day to see whether it works.
- The MC chain is almost complete. I succeeded at writing a second ART generator and having the ddt-filter executable decide which generator to run based on the provided FHICL parameter.
- Duyang will write a Wiki entry on how to create the raw millislice files from MC simulation.
- Brian would like a set of FD cosmic rays.
- Zukai also wants FD cosmic rays
- Jan wants Nue MC.
- I noted that we are not quite ready for large scale MC orders yet, but I would prefer to get the MC chain working first. If everything goes smoothly, we might be able to have some cosmic ray MC at the collaboration meeting.
- We are interested in the time it takes for each trigger module. I ran over ~10,000 NDOS events with the timing module turned on and then parsed the output using a simple shell script. When I showed the first analysis of these numbers we noticed very large maximum values, but it turns out that these large values occur in the first event only. Here are the time values for the first event:
HitProducer: 18.4 ms TDC Sort: 4.5 ms Time Slice: 4.4 ms Trigger Results: 0.2 ms Total Event: 28.4 ms
- The attached file shows the distributions for the next 10,000 events (i.e. excluding the first event).
- Andrew noted that we spin up the DAQ system before we actually start collecting data, so the delays of the first event may be irrelevant for us.
- Zukai would like his module to be profiled.
- Evan asked me to profile his various slicers as well.
Message to Global Trigger¶
- Andrew has not had a chance to speak with Marc yet.
Nue Studies and Clustering¶
- Jan is working on his Nue module and is taking the space slices from Evan. The module has a pass and fail writer, which seem to work fine.
- The firmware has been released by Harvard.
- Brian and Andrew agree that we should leave things as they are and not try to write our own SoftRelTools, so we should continue to use "addpkg_svn". Brian noted that we might move away from SRT in the future anyway. Lynn is working on getting Cmake to work with the nutools.
- We have three working group sessions on Thursday. Andrew will provide goals for this meeting.
- Andrew asked for people to showcase their work during the plenary sessions. Zukai and I volunteered for this.
- Andrew asked Duyang to generate a set of raw millislices for cosmic rays.
- I would like to show my profiling work.
- We started talking about deploying the DDT software on the FD cluster. I will get started with this, so we'll have to see how much progress can be made by next week.
- Andrew noted that we have DCMs connected up north and that they can run in simulation mode to generate some garbage output, which the buffer nodes can pick up and feed into our DDT chain.
- Yuri will look into expanding the DAQApplicationManager for the FD.
- Gavin will try to provide some nice EVDs.
- Gavin needs a DDT tag. We agreed to call it R(date) (e.g. R13.01.14). He mentioned that he would need all of the different build flavors (i.e. maxopt, prof, debug). We'll sort the details of this out offline.