Difference between revisions of "EPSCI Group Meeting Oct. 19, 2020"

From epsciwiki
Jump to navigation Jump to search
(Created page with " The meeting time is 10:00am. === Connection Info: === <div class="toccolours mw-collapsible mw-collapsed"> You can connect using [https://bluejeans.com/253300597 BlueJeans V...")
 
 
Line 62: Line 62:
 
<hr>
 
<hr>
 
=== Minutes: ===
 
=== Minutes: ===
Attendees: David L., Carl T., Nathan B., Thomas B., Kishan R.
+
Attendees: David L., Carl T., Nathan B., Thomas B., Kishan R., Vardan G.
  
'''Announcements'''
+
'''Fortnight Meeting'''
* Carl called in from his vacation up in Michigan
+
* Coding standards discussion next week. Read up!
* This is no news to report from Michigan.
 
  
'''Fortnight Meeting'''
+
'''RT2020'''
* Paper chosen for fortnight discussion was more of a usage manual than scientific paper. Those who had not read it already were not encouraged to do so. This coupled with not having a Code Review topic for today we decided to cancel the Fortnight paper/Code Review meeting for this afternoon and instead meet in 2 weeks to discuss the [https://guide.esciencecenter.nl/#/ coding standards] Nathan pointed us to some weeks ago.
+
* Vardan attending
 +
** Operational vs. Informational Technology
 +
** Lots of efforts to put intelligence closer to detectors
 +
** MLS4ML - CERN tool to convert models from most common packages to form suitable for accelerators (e.g. GPUS)
  
 
'''EVIO-6'''
 
'''EVIO-6'''
 +
* Carl out of town last week
 
* Documentation
 
* Documentation
** Tested use of [https://pandoc.org/ pandoc] to convert existing code into sphinx markdown for use with [https://readthedocs.org/ ReadTheDocs].
+
** Will work on integrating Doxygen comments into the produced ReadTheDocs pages this week.
** Numerous issues with the conversion that still required a lot of manual tweaking
 
** Doxygen comments not incorporated yet into the produced ReadTheDocs pages.
 
  
 
'''JANA2'''
 
'''JANA2'''
* HDDM event source working (used for GlueX MC data)
+
* EVIO event sources working
* Working on Plugins/Processors
+
* Working on hd_dump and hd_root to start consistency checks with original JANA
  
 
'''A.I.'''
 
'''A.I.'''
 
* JLab AI website
 
* JLab AI website
** Kishan completed Drupal training on Friday and now has access as an editor (David is still pubisher)
+
** Some changes submitted by Kishan. (David still needs to approve)
** Plan is to implement proposed updates over the next 2 weeks.
 
 
* Hydra
 
* Hydra
** Training script ported to Tensorflow 2.3 and is working
+
** Kishan restructuring predict script to use classes and make it more flexible for 3rd party users
** Some discussion of other features that could be added, but that are not considered the highest priority at the moment
+
** Goal to have a paper written this year
** Kishan is working on porting the predict script to TF2.3
+
*** Working on a publishable form of the code by then so it can be referenced
** Paper
+
*** Thomas has almost completed a draft of the introduction in Overleaf
*** Goal is to have complete draft of first Hydra paper by Dec. 2020
 
*** An empty Overleaf document has been created
 
** Thomas has started designing a general use AI report class in Python
 
** Kishan noted most recent Tensorflow version available via conda is 2.2
 
*** David noted Tensorflow 2.3.1 was available via venv (will send Kishan link to documentation on installing venv as user-space kernel in jupyterhub.
 
  
 
'''SRO'''
 
'''SRO'''
* Vardan is attending the RealTime conference this week and was not at the EPSCI meeting today. He is presenting a poster on ERSAP. The 1 minute introduction is to be shown today at 12:45. (Registration to RT2020 is free)
+
* Vardan is porting many of the streaming microservices to C++
 +
* Will consult with Carl as needed since given his recent experience with related algorithms such as highly efficient ring buffers.
 +
* Vardan has provided some slides needed for Marco's talk on Thursday (David needs to supply some as well.)
 +
 
 +
'''DAQ'''
 +
* Vardan enhancing the filtering feature of the CODA Message Browser at the request of Bryan Moffit. He needs it to help debug some issues in Hall-A/C.
  
 
'''SPACK'''
 
'''SPACK'''
* Thomas discovered a feature that makes it easy to generate a skeleton configuration for a package using only a URL to something like a github repository
+
* No time to work on this during the past week
* Five of the packages we wish to support are not currently part of the standard SPACK package set
+
 
* A rich dependency handling system is included, but will need a little time to figure out the best way to use it at JLab
+
'''MCWrapper'''
 +
* Thomas has integrated support for JEF (JLab Eta Factory experiment) into MCWrapper at their request.
 +
* GlueX has requested feature to add a custom JANA config file to MCWrapper jobs.
  
 
'''Offsite Computing'''
 
'''Offsite Computing'''
 
* IU - Big Red3
 
* IU - Big Red3
** First job run manually on Big Red3
+
** Initial SWIF2 integration complete, though still working out issues with getting first job to run
** Working to get integrated into swif2 this week (possibly)
 
* PSC
 
** Early User Program application submitted for Bridges-2. (will only last 1 month in Jan.)
 
 
* NERSC
 
* NERSC
** AY21 application submitted
+
** Met with NERSC analyst, Chris L.;, and Jie C. on Friday to discuss multi-job jobs. Better understanding of problem and possible solutions. We now have a plan to work towards a solution.
 
* OSG
 
* OSG
** Thomas and Wes met with Edgar last week to go over more details. Wes presented the current list of issues/questions
+
** Wes was told by Edgar to open a ticket once things were set up on the JLab end. This has been done, but no further word yet.
** Many of the needed pieces are now set up and configured.
+
** Possible first jobs could come in this week to run on the 4 dedicated OSG-dev nodes. Depends on schedule of Wes and Edgar when it will be completely ready.
** On target for having it run by end of year. (may actually be sooner, but no date offered at meeting)
 
** MCWrapper
 
*** JEF program in Hall-D has expressed interest in adding a variation for them to the Hall-D implementation
 
*** Something was set up for SoLID some time ago, but Thomas has not heard from the one person who was going to use it for some time. He will reach out again to see if there are any issues.
 

Latest revision as of 16:01, 19 October 2020

The meeting time is 10:00am.

Connection Info:

You can connect using BlueJeans Video conferencing (ID: 253 300 597). (Click "Expand" to the right for details -->):

Meeting URL
 https://bluejeans.com/253300597?src=join_info

Meeting ID
253 300 597

Want to dial in from a phone?

Dial one of the following numbers:
+1.888.240.2560 (US Toll Free)
(see all numbers - https://www.bluejeans.com/premium-numbers)

Enter the meeting ID and passcode followed by #

Connecting from a room system?
Dial: bjn.vc or 199.48.152.152 and enter your meeting ID & passcode

Agenda:

  1. Previous meeting
  2. Announcements
  3. Real Time 2020
  4. SoLID Software Status
  5. Ongoing Activities
    • EVIO-6 (Carl)
    • JANA2 (Nathan)
      • GlueX port
    • A.I.
      • JLab AI webpage (Kishan)
      • Hydra (Kishan/Thomas)
      • Experimental Controls (David)
    • SRO (Vardan)
      • ERSAP
      • Hall-B/D TriDAS
    • JLab Common Environment (CE) + SPACK (Thomas)
    • Offsite Computing
      • BigRed3 at IU (David)
      • NERSC (David)
      • OSG (Thomas)
  6. GUI for Calorimeter calibration scripts (Hall-D Request) (Thomas)
  7. Publications
  8. Coding Standards
  9. AOT



Minutes:

Attendees: David L., Carl T., Nathan B., Thomas B., Kishan R., Vardan G.

Fortnight Meeting

  • Coding standards discussion next week. Read up!

RT2020

  • Vardan attending
    • Operational vs. Informational Technology
    • Lots of efforts to put intelligence closer to detectors
    • MLS4ML - CERN tool to convert models from most common packages to form suitable for accelerators (e.g. GPUS)

EVIO-6

  • Carl out of town last week
  • Documentation
    • Will work on integrating Doxygen comments into the produced ReadTheDocs pages this week.

JANA2

  • EVIO event sources working
  • Working on hd_dump and hd_root to start consistency checks with original JANA

A.I.

  • JLab AI website
    • Some changes submitted by Kishan. (David still needs to approve)
  • Hydra
    • Kishan restructuring predict script to use classes and make it more flexible for 3rd party users
    • Goal to have a paper written this year
      • Working on a publishable form of the code by then so it can be referenced
      • Thomas has almost completed a draft of the introduction in Overleaf

SRO

  • Vardan is porting many of the streaming microservices to C++
  • Will consult with Carl as needed since given his recent experience with related algorithms such as highly efficient ring buffers.
  • Vardan has provided some slides needed for Marco's talk on Thursday (David needs to supply some as well.)

DAQ

  • Vardan enhancing the filtering feature of the CODA Message Browser at the request of Bryan Moffit. He needs it to help debug some issues in Hall-A/C.

SPACK

  • No time to work on this during the past week

MCWrapper

  • Thomas has integrated support for JEF (JLab Eta Factory experiment) into MCWrapper at their request.
  • GlueX has requested feature to add a custom JANA config file to MCWrapper jobs.

Offsite Computing

  • IU - Big Red3
    • Initial SWIF2 integration complete, though still working out issues with getting first job to run
  • NERSC
    • Met with NERSC analyst, Chris L.;, and Jie C. on Friday to discuss multi-job jobs. Better understanding of problem and possible solutions. We now have a plan to work towards a solution.
  • OSG
    • Wes was told by Edgar to open a ticket once things were set up on the JLab end. This has been done, but no further word yet.
    • Possible first jobs could come in this week to run on the 4 dedicated OSG-dev nodes. Depends on schedule of Wes and Edgar when it will be completely ready.