AWS meeting 2023 04 13

EESSI/AWS sync meeting (2023-04-13)

• link to AWS project doc: https://docs.google.com/document/d/1CHG9fCh2LkfJ-EI8J-_Wr5NpHL5iwm8Wu6syfK9h7-c

Next meeting

• Thu 11 May 2023, 12:00 UTC
• Thu 8 June 2023, 12:00 UTC

Notes previous meetings

• 9 Mar 2023: https://github.com/EESSI/meetings/wiki/AWS-meeting-2023-03-09
• 11 Jan 2023: https://github.com/EESSI/meetings/wiki/AWS-meeting-2023-01-11

Notes 13 April 2023 March 2023

• status update on sponsored credits
  • see slide 14 in https://raw.githubusercontent.com/EESSI/meetings/main/meetings/EESSI_meeting_20230406.pdf
  • ~$2.6k credits consumed in Mar'23
    • significant increase compared to previous months
  • remaining credits (~$7.1k) should be sufficient until end of July'23 (at current burning rate)
  • budget alert are set at $1k + $2k / month
  • to be reviewed in May, planning for refresh in June
• MultiXscale EuroHPC project is now on the rails
  • kickoff meeting was on 20-23 March'23
    • see https://www.multixscale.eu/wp-content/uploads/2023/04/MultiXscale-Kick-off-meeting_Press-Release_vf.pdf
    • poster at EuroHPC Summit: https://www.multixscale.eu/wp-content/uploads/2023/03/32-Poster-MultiXscale.pdf
      • which questions were raised by people visiting the poster?
    • CASTIEL2 coordination project for EuroHPC CoEs and NCCs
• primary goals for EESSI in coming weeks/months
  • new EESSI pilot version (2023.04)
    • slowly getting past problems that presented themselves in bootstrapping Gentoo Prefix
  • start using build-and-deploy bot for all software builds (software layer + compat layer)
    • see https://github.com/EESSI/eessi-bot-software-layer
  • extend software stack
    • more recent toolchains
    • apps relevant for MultiXscale (ESPResSo, waLBerla, LAMMPS), and other EuroHPC CoE's (GROMACS, OpenFOAM, ...)
    • eye catchers like AlphaFold, OpenFold, ...
  • NVIDIA GPU support
    • see https://github.com/EESSI/software-layer/pull/172
    • not knowing which GPU driver version is available implies possible need for CUDA compat libraries
      • compat libraries are installed into /opt/eessi, which gets symlinked into EESSI software stack
      • write permissions to /opt/eessi needed
  • EESSI test suite
    • https://github.com/EESSI/test-suite, using ReFrame
    • focus on portability of test suite + performance
    • currently only a GROMACS test, which will serve as a blueprint for other tests
    • working on tests for OSU Microbenchmarks, TensorFlow, ...
  • Add support for customising EESSI initialisation (including enabling tracking support)
    • different variants of EESSI init script can be provided
    • sites (like AWS) providing EESSI can opt-in to using an init script that does usage tracking (by leveraging "variant symlink" feature in CVMFS)
    • /cvmfs/pilot.eessi-hpc.org/latest/init/bash -> /cvmfs/pilot.eessi-hpc.org/latest/init/bash_tracking_aws
    • AWS-specific init script can also set additional environment variables required in AWS environment (related to EFA for example, etc.)
  • tutorials
    • EESSI introductory tutorial at HPC Knowledge Meeting 2023 in Barcelona (17-18 May 2023 - https://hpckp.org/annual-meeting)
    • "CernVM-FS Best Practices" online tutorial
      • fall 2023
      • in collaboration with CernVM-FS developers
      • focus on use of CernVM-FS on HPC systems
      • we hope that EESSI configuration is included in CernVM-FS release by then (one step less to deploy EESSI)
• ISC'23
  • AWS presence?
    • workshop/tutorial on Sunday: Best Practices for HPC in the Cloud
      • maybe mention EESSI alongside Spack w.r.t. software deployment?
    • HPC Tech Shorts interviews
      • example interview: https://youtu.be/o_noqdK5Sdc
      • get Brendan in touch with Eli (HPCNow!) to see if an on-floor interview is possible
        • "How EESSI helped us solve X"
        • obvious call to action at the end
        • EESSI as streaming software pitch (a la Spotify/Netflix)
    • walk-in meeting rooms (scheduled)
  • additional opportunities for EESSI promotion?
    • booth talk at MS Azure booth
    • MultiXscale will have presence in EuroHPC booth (poster, presentation time)
• Q&A
  • container images
    • container images to archive deprecated EESSI versions
    • service to create container images on-demand?
    • see container image for accessing EESSI
      • https://eessi.github.io/docs/pilot/#accessing-the-eessi-pilot-repository-through-singularity
      • https://eessi.github.io/docs/getting_access/eessi_container/
    • can also create a squashfs file from CernVM-FS and mount that in container
    • think about long-lived clusters vs short-lived (AWS Batch), persistent storage, ...
    • need to check how much data is being pulled in when using EESSI pilot container (incl. container image itself)
  • ParallelCluster repositories for accessing software modules
    • incl. Spack, could also cover EESSI
    • similar to what done for Cluster-in-the-Cloud, azhop (Azure), Magic Castle (https://github.com/ComputeCanada/magic_castle)