SIMD-MIMD Real-Time Comparisons (Our Chapter 7)

1
SIMD-MIMD Real-Time Comparisons(Our Chapter 7)

• References
• Stankovic, Spuri, Ramamritham, Buttazzo,
  Deadline Scheduling for Real-Time Systems,
  Kluwer, 1998, ISBN 0-7923-8269-2.
• Stankovic, Spuri, Natale, Buttazzo, Implications
  of Classical Scheduling Results for Real-Time
  Systems, IEEE Computer, Vol. 28, No 6, pp.
  16-25, June 1995.
• Meilander, Jin, Baker, Tractable Real-Time Air
  Traffic Control Automation, Fourteenth IASTED
  International Conference on Parallel and
  Distributed Computing and Systems (PDCS02), pp.
  483-488, November 2002.
• Meilander, Baker, Jin, Importance of SIMD
  Computation Reconsidered, working copy handout.
• Initially we will go through most of the slides
  used for the presentation of the paper,
  Tractable Real-Time Air Traffic Control
  Automation.
• Presentation given by Will Meilander at PDCS
  Conference in November 2002.
• The PDCS slides will follow this set of slides.
• A polynomial time MP solution to the ATC is not
  believed possible. In particular,
• MP solutions to virtually all real-time problems
  today include an online solution to one or more
  dynamic scheduling problems.
• Dynamic scheduling is an NP-hard problem.
• The MP cannot solve most real-time problems using
  a static scheduling algorithm (to avoid using
  dynamic scheduling).

2

• As observed in the earlier set of slides, a low
  order polynomial time solution to the ATC problem
  is possible on the AP.
• Open Question Does there exist a splitting
  problem that has a polynomial time solution
  using an AP but all solutions using an MP involve
  solving an additional problem that is NP-hard.
• Additional restrictions can be made, if needed.
• Such a result would establish that MPs are not
  more powerful than SIMDs, as is currently
  believed by many professionals in the area.
• This work raises other open questions as well,
  both practical and theoretical.
• How much can one move away from the SIMD
  architectures and programming style towards MIMD
  before encountering the MIMD difficulty
• That is, of having to solve NP-hard problems
  in order to solve original problem
• In particular, how can the multiple instruction
  streams for MASC interact so as to avoid this
  MIMD difficulty?