Basic HTML version of Foils prepared August 7 98

Foil 6 Performance of HPF

From DoD HPF Training -- 1. Introduction to HPF DoD Training and Others -- 1995-98. by Chuck Koelbel -- Rice University
1 Performance is highly dependent on the compiler and on the nature of the code
  • Best performance is for data-parallel computations
  • Tie ³natural² parallelism in the problem to distributed array dimensions
2 Commercial compilers are now competitive with MPI for regular problems
  • Subscripts like A(L+1,K-1)
3 Research continues on irregular problems and task parallelism
  • Subscripts like A(INDX(I1,2))
  • Tree searches with branch-and-bound cutoffs
in Table To:

© 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 Aug 9 1998