Computing R&D Micro-Retreat

US/Central
CDF Big Room (Bldg 327)

CDF Big Room

Bldg 327

Adam Lyon (Fermilab), Jim Kowalkowski (Fermilab)
Description

Zoom Info will be e-mailed

The agenda is done. Please check your talk and upload slides. Let Adam ([email protected]) know if you think your talk is in the wrong place. Some talks are grouped (e.g. Frameworks and CMS directions). Please upload slides as one file or several into that slot. 

Holy Cow! We have a lot of talks! Please try to keep your talks short. You will need to stop when you time is up. 

 

Some of you are aware that on April 26, there is an All-Scientists' Retreat where the topic will be "What are the interests of the Fermilab scientists for the decade or so following 2026". Jim Kowalkowski and I are heading up the Computational Science track. We need your input for what to present at this all-scientists' retreat, and so we are having a "Computing R&D Micro-Retreat."

Predicting what Computing will look like in 2026-2036 is a very difficult task ... and we're not going to try. Instead, we would like to talk about Computing R&D steps we may want to perform from now through 2026 or so to get us into that era. We clearly cannot cover all aspects of computing and we need to stay at a big picture scale and think far future. Here are some things that may help the discussion...

Jim and I identified three paths for how "new stuff" comes about in Computing...

  1. We receive requirements from experiments on upcoming needs to be met.
  2.  Forward thinking from us (e.g. things we know are coming that the experiments don't know to ask for, such as Exascale, Quantum Computing, SciDAC topics, etc)
  3. We react to new topics and techniques that interest SCD and other division scientists. I think the explosion in machine learning is a great example of this path.

It's impossible to predict path 3 by definition. We will try to get input via path 1 by hearing from people in the other retreat tracks (I'll invite them to our meeting) and any ideas from SCD people. Though I think we'll expect that articulating specific computing needs for experiments over 10 years out will be difficult. We'll identify people in computing who can speak to path 2.

We identify three broad core areas for R&D...

A) Computational Software (frameworks, algorithms, simulation, analysis, adapting to new architectures)

B) Operating Computing Systems (Storage, Networking, Load management, Workflows, Middleware, System management and Engineering, connecting to grids and clouds, e.g. HEPCloud)

C) DAQ R&D (triggers, firmware, front/back ends, drivers, communications, electronics)

NEW: A recording of the meeitng is available below. Clicking on the video should download it to your computer (it's about 150 MB). If instead you get a playback screen that doesn't work, right-click on the screen and choose "Download Video" or something similar. 

Video
    • 13:30 13:40
      Introduction
      • 13:30
        Why are we here? 10m
        Speaker: Dr Adam Lyon (Fermilab)
        Slides
    • 13:40 14:55
      Stepping stones towards DUNE/HL-LHC (near and farther future)
      • 13:40
        CWP and S2I2 5m
        Speaker: Dr Rob Kutschke (Fermilab)
        Slides
      • 13:45
        HEPCloud Direction 5m
        Speaker: Dr Burt Holzman (FNAL)
        Slides
      • 13:50
        Machine Learning Directions 5m
        Speaker: Dr Brian Nord (Fermilab)
        Slides
      • 13:55
        ROOT i/o directions 5m
        Speaker: Mr Philippe Canal (FERMILAB)
        Slides
      • 14:00
        Reconstruction directions 5m
        Speaker: Giuseppe Cerati (Fermilab)
        Slides
      • 14:05
        Simulation Directions 5m
        Speaker: Soon Yung Jun (Fermilab)
        Slides
      • 14:10
        CMS Directions 15m
        Speakers: Dr Christopher Jones (Fermilab), Nhan Tran (FNAL), Dr Oliver Gutsche (Fermi National Accelerator Laboratory)
        1. CMS R&D Overview (PDF)
        Additional Material for CMS R&D Overview
        Slides
      • 14:25
        Framework Directions 20m
        Speakers: Mr Jim Kowalkowski (Fermilab), Dr Kyle Knoepfel (Fermilab), Dr Marc Paterno (Fermilab), Saba Sehrish (Fermilab)
      • 14:45
        Heterogenous Computing and Coherent Interconnects 5m
        Speaker: Dr Michael Wang (Fermilab)
        Slides
      • 14:50
        Cosmic Software Directions 5m
        Speaker: Dr James Annis (Fermilab)
        Slides
    • 14:55 15:30
      Problems to Confront (~2026+)
      • 14:55
        Post Moore Introduction 10m
        Speaker: Mr Jim Kowalkowski (Fermilab)
        Slides
      • 15:05
        Computing Security 5m
        Speaker: Dr Mine Altunay (FNAL)
      • 15:10
        DAQ Directions 5m
        Speaker: Dr Gustavo Cancelo (fermilab)
        Slides
      • 15:15
        TDAQ Workshop Summary 5m
        Speaker: Mr Alan Prosser (Fermilab)
        Slides
      • 15:20
        Energy Frontier Directions 5m
        Speaker: Dmitri Denisov (Fermilab)
        Slides
      • 15:25
        More for CMS 5m
        Speaker: Lindsey Gray (Fermilab)
        Slides
    • 15:30 15:40
      Discussion: Core areas. Practices. CERN & NSF