PYTHIA: a knowledge-based system to select scientific algorithms

  • Authors:

  • Sanjiva Weerawarana;Elias N. Houstis;John R. Rice;Anupam Joshi;Catherine E. Houstis

  • Affiliations:

  • Purdue Univ., West Lafayette, IN;Purdue Univ., West Lafayette, IN;Purdue Univ., West Lafayette, IN;Univ. of Missouri, Columbia;Univ. of Crete, Heraklion, Greece

  • Venue:
  • ACM Transactions on Mathematical Software (TOMS)
  • Year:
  • 1996

Quantified Score

Hi-index 0.02

Visualization

Abstract

Problem-solving environments (PSEs) interact with theuser in a language “natural” to the associated discipline,and they provide a high-level abstraction of the underlying,computationally complex model. The knowledge-based system PYTHIAaddresses the problem of (parameter, algorithm) pair selection within ascientific computing domain assuming some minimum user-specifiedcomputational objectives and some characteristics of the given problem.PYTHIA's framework and methodology are general and applicable to anyclass of scientific problems and solvers. PYTHIA is applied in thecontext of Parallel ELLPACK where there are many alternatives for thenumerical solution of elliptic partial differential equations (PDEs).PYTHIA matches the characteristics of the given problem with those ofPDEs in an existing problem population and then uses performanceprofiles of the various solvers to select the appropriate method givenuser-specified error and solution time bounds. The profiles areautomatically generated for each solver of the Parallel ELLPACKlibrary.—Authors' Abstract