On Sat, 31 Jul 2004, Alastair Reid wrote:
On Saturday 31 July 2004 09:11, Matthew Roberts wrote:
tell us the fundamental reason it is hard to compile Haskell for multiple processor machines - in broad strokes.
Unpredictability, heavy use of heap and overly fine granularity of inherent parallelism.
How well does GPH solve the first two of these problems (I think it solves the third by asking the programmer for parallel annotations)? Also, if you use forkOS, doesn't your app inevitably distribute itself accross multiple processors if the OS runs threads that way? If the OS does run threads that way, are you saying that forkOS is *very* inefficient (or does the forkOS implementation in e.g. GHC solve the problems you describe)? -Alex- ______________________________________________________________ S. Alexander Jacobson tel:917-770-6565 http://alexjacobson.com