Supercomputers for Solving PDE Problems
Abstract
Researchers investigated several supercomputer architectures in the context of assessing their performance in solving PDE problems. Main results are: assessed various classes of paralled and vector supercomputers and parallel software issues toward solving PDE problems; developed a multipipeline networking technique for compound vector processing, developed an orthogonal multiprocessor for large-grain scientific computations; improved parallel efficiency of a domain decomposition method, the DD algorithm. Keywords: Computer systems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1988
- Accession Number
- ADA204611
Entities
People
- Kai Hwang
Organizations
- University of Southern California