Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Research Computing Seminar

Research Computing Seminar

A discussion of the research computing facilities available at ILRI, Kenya.

Avatar for Alan Orth

Alan Orth

June 06, 2013
Tweet

More Decks by Alan Orth

Other Decks in Technology

Transcript

  1. 1. Users! 2. Potential users! 3. Managers of users! The

    slides aren't technical, but I can't promise I won't start speaking in tongues... Intended audience
  2. Harder, Better, Faster, Stronger Work it (harder) Make it (better)

    Do it (faster) Makes us (stronger) -- Daft Punk Research-oriented computing is different than corporate ICT!
  3. History - Few users with "heavy" computing / data requirements

    (bioinformatics + some GIS) - Etienne de Villiers and the "HPC" (lots of BLAST!) - Other users simply ran analyses on their workstations (small data sets, "first-gen" sequence data, statistics etc)
  4. Things are changing - New platforms (high throughput on all

    fronts!) - New collaborators (with new toys!) - More users - More projects - Different kinds of data - etc...
  5. Play nice, use SLURM! - Simple Linux Utility for Resource

    Management - Everyone is allowed to use a "fair" amount of resources - If you go over, you have to "queue"
  6. Why SLURM? - How can I run structure for populations

    1.. 10, and repeat each run 3 times? - How can I run newbler for 1 week?! I have to shut my computer down! - How can I get an email when my pipeline finishes? - I want to run R with 4 CPUs right now!
  7. Software available - mira, newbler, velvet, trinity - BLAST, bowtie,

    bwa - figtree, Dendroscope, MEGAN, PhyML - mrbayes, bayescan, R - structure, admixture, apt - CLC Genomics Workbench (kitchen sink) More in /export/apps/*!
  8. Proposal: HPC Users Group - Meet once a month -

    Promote community - Share scripts / pipelines / experiences - Yours truly as fearless leader - Promote love of Linux and coffee (?)