Codemash 2018 – A Brief Introduction Into Concurrent Programming

Presenter: Rob Keefer

• Example of processor concurrency
  • 1 hertz – 1 ‘turn of the crank’ per second
    • modern cpu’s 300 Gigahertz – 3 billion per second
  • Chess player playing multiple games example
    • Player looks to see if player has moved yet, if not moves on to different game
      • Processor does this as well, looks at queue, if nothing happening moves to next thread.
    • Dual core, similar to having two players servicing many one on one matches.
      • Sharing resources means don’t get one for one throughput improvement when adding another core.
  • Problems with concurrency
    • starting processes
    • communicating between processes
    • Shared Resources
      • Live version of five philosophers blocking problem
        • Waiting for two chopsticks
        • Deadlock  as shown by Dr. Suess Zaks video
      • Livelock
        • Two people meeting in a hallway, keep blocking by making same moves to mitigate. Still active, so live lock
      • Race Condition
  • Approaches to managing resource contention
    • Traditional – preventive, avoid failure
      • Monitor
      • Mutex
        • Double lock, one to lock resource and one to make sure we are only one with the inner lock.
      • Prioritization Algorithms
    • Newer approaces – embrace failure
      • embrace process death and ressurection
      • support atomic functions (then no need to interleave operations)