Basic HTML version of Foils prepared May 18 99

Foil 14 Numerical Computing in Java

From Java Grande Overview IPPS/SPDP 1999 International Workshop on Java for Parallel and Distributed Computing San Juan Puerto Rico -- April 12 99. by Geoffrey C. Fox


1 So Java not only will run anywhere but can be expected to get same answers everywhere
  • As long as use of Native (C, Fortran) libraries and indeterminate concurrency controlled
2 Natural tension between performance (both in terms of speed and precision) and reproducibility
  • Need to encourage more research into "reproducible algorithms" but support both goals
3 Java has particularly bad floating point performance due to
  • As in all applications, immature compiler (JIT) technology
  • does not exploit proprietary hardware features (extended precision, fused multiply add)
  • does not allow well known compiler optimizations (associativity)
4 Solution requires "Change in Java Rules" and better compilers

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 Tue May 18 1999