Basic HTML version of Foils prepared 15 March 1996

Foil 9 HPF computational model

From Complex Systems and Parallel Computing CPSP713 Case studies in Computational Science -- Spring Semester 1996. by Geoffrey C. Fox
The underlying intuition behind HPF computational model is that
  • an operation on two or more data object is likely to be carried out much faster if they all reside on the same processor
  • it may be possible to carry out many such operations concurrently if they can be performed on different processors
The main mechanism for achieving parallelism in HPF is the data partition annotation of HPF. Distribution of control is derived from the data distribution, using the owner compute rule: each assignment is executed by the processor(s) that own the assigned variable. (However, a compiler may relax the owner rule, in order to improve performance.)

Northeast Parallel Architectures Center, Syracuse University, npac@npac.syr.edu

If you have any comments about this server, send e-mail to webmaster@npac.syr.edu.

Page produced by wwwfoil on Sun Feb 22 1998