Block-Oriented, Local-Memory Linear Equation Solution on the CRAY-2. Part 1. Uniprocessor Algorithms.
MICHIGAN UNIV ANN ARBOR SUPERCOMPUTER ALGORITHM RESEARCH LAB
Pagination or Media Count:
Experience with the CRAY-2 on the effects of common memory speed and loading on performance indicate that local-memory-based algorithms have potentially a large advantage. The performance of a number of common- and local-memory algorithms are compared for the LU factorization of a dense systems of equations on the CRAY-2. Results of both Fortran and assembly language implementations are given. Author
- Computer Programming and Software
- Computer Hardware