Issues in And-Parallelism - Independence

• Correctness generally OK - efficiency issues:

  

  

  p(a).
  q(b) :- large computation.
  q(a).
  

  • Parallelism always correct (pure programs).
  • More work can be done (speculativeness).
  • Worse: ``no-slowdown'' not preserved.
  Some aspects new and particular to languages with non-determinism!

• Notions of goal independence proposed to avoid such problems.

• Informal notion: a computation ``does not affect'' another
  (also referred to as ``stability'').

• Detection of independence:
  • Run-time (a-priori conditions) [Conery, Lin, AKL].
  • Compile-time [Chang].
  • Mixed: conditional execution graph expressions [DeGroot, Hermenegildo].
  • User controlled: explicit parallelization (concurrent languages).

• Granularity considerations (see later)