• Proj 2 due today
    • Will accept something a little late
  • Skipping project 3
  • HW 3: final proposal (due next Tues)

Final projects

  • Goal: Understand/improve performance on code of interest
    • Could be single-core tuning
    • Could be parallelization
    • Could be an algorithm investigation
    • Or more! Just ask!
  • Great if it is applicable to your research
    • Also fine if it is completely different
  • Main outcome is a report
    • Great if it is at a conference paper level
    • But this is not necessary for success
  • Teams of 1-3, your choosing
  • Idea: advertise your project idea on Piazza
  • If you want ideas, ask me
  • Can also look at SuperComputing 20