May 15, 2020 JLabCE Management + PSSC

From epsciwiki
Jump to navigation Jump to search

Meeting to discuss EPSCI maintenance of JLabCE + Resurrection of Physics Software Support Committee

Connection Info:

You can connect using BlueJeans Video conferencing (ID :740 025 234). (Click "Expand" to the right for details -->):

Meeting URL
 https://bluejeans.com/740025234

Meeting ID
740 025 234

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


Reference

Minutes

Attendees: David L., Thomas B., Brad S., Dmitry R., Mark I., Mauri U., Nathan Baltzell., Ole H., Zhiwen Z.

JLabCE Introduction

  • Mauri was put on the spot with no warning to give a brief introduction to JLabCE.
    • System for building and maintaining versions of dependency packages used by various groups throughout the lab
    • Copies of source code kept at JLab
      • Allows permanent, stable availability
      • Allows for custom, local patches (e.g. GEANT4)
    • Binaries maintained in /site/12gev_phys with variations for version number, OS, and compiler
    • Docker container maintained with a standard version set
    • Suggested we consider CVMFS distribution

What do Halls currently use

  • Hall-A
    • Main thing used is ROOT build
    • Analyzer is not part of CE so is built/maintained external to it
    • Many people use private installations of dependency packages (may not be aware of JLabCE versions)
    • CREX and PREX use specialized pairity analyzer software (may not benefit much from centralized management)
  • Hall-B
    • Uses some packages built as part of CE, but environment set by hall-specific Modules config files
  • Hall-C
    • Similar to Hall-A in that experimental groups tend to be loosely coupled so not much formal interdependency on software
    • Some groups migrating more to containers (Singularity and Docker)
  • Hall-D
    • Version sets maintained in XML files
    • Single XML file is used to build packages, set environment, and document a complete version set
    • No dependency on JLabCE at all
  • Hall EIC
    • Used home-grown ejpm up to now
    • Looking to transition to spack

Modules

  • Nathan currently using in Hall-B with good success
  • Well established standard
  • Supported in spack
  • Allows for "uber"-modules to group multiple packages together

spack

  • Dmitry suggested we take a strong look at spack which provides the functionality we need in a modern, popular package manager
  • General consensus we should look into this
  • David volunteered himself and Thomas to investigate spack and to develop a proposal in consultation with Mauri and Dmitry
  • Once a proposal is formed, another ad hoc meeting will be called to review and refine it.