GlueX Software Meeting, November 10, 2020
GlueX Software Meeting
Tuesday, November 10, 2020
3:00 pm EST
BlueJeans: 968 592 007
- Review of Minutes from the Last Software Meeting (all)
- Review of recent issues and pull requests:
- Review of recent discussion on the GlueX Software Help List (all)
- Action Item Review (all)
Present: Alex Austregesilo, Sean Dobbs, Mark Ito (chair), Naomi Jarvis, David Lawrence, Simon Taylor, Nilanga Wickramaarachchi, Beni Zihlmann
There is a recording of his meeting on the BlueJeans site. Use your JLab credentials to authenticate.
- work disk full again. We are still very limited on space. Only 14 TB free (6%). See this plot of current usage.
- New release of Build Scripts: version 2.3. Sets us up to have builds on more advanced distributions, e.g., CentOS 8.
- New git repository: hd_singularity. New repository with scripts for building Singularity containers.
- New version set: 4.30.0 Update version set.
Review of Minutes from the Last Software Meeting
We went over minutes from the October 13th meeting.
Scientific Computing Hardware
Mark reported some news from SciComp.
- The tape library upgrade last week went well. After two libraries were converted to one all tapes are loadable into all drives.
- The Computer Center is very close to allowing jobs from the OSG to run on the JLab farm.
- The work disk space will be supplemented by 2018-era file servers in the not-too-distant future.
- More farm nodes are being purchased, perhaps with the new AMD Milan architecture. It should come to about 80 more nodes.
Future Trends in Nuclear Physics Computing
Mark mentioned that the organizing committee of the Future Trends Workshop released a draft report at the Round Table Meeting this morning. Comments on the draft are welcome. Find it here.
Mark pointed out the package that Dmitry Romanov is using for unit test for CCDB. It is called Catch2.
Review of recent issues and pull requests
- halld_recon issue #462 TOF dE/dx written as all zeros in monitoring hists and ROOT trees since 2020-07-10. In response to Andrew Schick's nice problem report, Beni has submitted a fix. The problem was the result of the introduction of dEdx1 and dEdx2 in addition to dEdx. dEdx got lost on the way to REST.
- Sean pointed out that the automatic builds, triggered by pull requests, have stopped working due to the input raw data file having been deleted from the work disk. [Added in press: Mark looked and discovered that the clean up of the pull request test area on the work disk was targeting the raw data file. That has been fixed now.]
- halld_recon issue #464 Issue with code for chi square comparison of two trees. Nilanga submitted this issue just before the meeting. Alex is looking at it. He sees a bug but it is not obvious why Nilanga sees crashes. Nilanga reported that the crash does not occur for all files.
- halld_sim issue #169 Amplitude for MC generation using a histogram in a ROOT file Nilanga submitted a request for a general method for specifying an amplitude using a histogram and generating events accordingly. Justin Stevens is working on this.
- halld_sim issue #168 resonance seeds in gen_amp causing inefficiency. Matt Shepherd identified a feature in gen_amp that concentrates events around 1 GeV when no resonances are specified. This is a problem when higher masses are desired. Alex reported that Rebecca Barsotti is working on a solution.
Copied from the last meeting:
- Ask the Computer Center about staging the start of jobs in time. (Mark)
- Make sure that the automatic tests of HDGeant4 pull requests have been fully implemented. (Mark, Sean)
- Finish conversion of halld_recon to use JANA2. (Nathan)
- Ping Dmitry on CCDB Issue #68. (Mark)
- Remind folks about their problematic cache files. (Alex)
- Finish CentOS 8 compatibility project. (Mark)