Basic HTML version of Foils prepared 15 March 1996

Foil 49 General Issues for Physical Optimization in Computing

From Physical Optimization and Physical Computation CPSP713 Case studies in Computational Science -- Spring Semester 1996. by Geoffrey C. Fox
Various optimizations are possible at each stage of mapping. These can be (but usually are not) formulated as minimizing
  • E = E1 (Performance) + E2 (Constraints)
This is a difficulty if constraint ensures correctness of calculation as in code generation phase of optimizing compiler for a digital computer. Then we must have E2=0 at T=0.
This "correctness" issue might not be present in generation of "programs" for neural (as opposed to digital) supercomputers.
We can study use of physical optimization in generation or execution of program

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