Timing Figures for Inverting Large Matrices Using the Staran Associative Processor.
Phase rept. Jan-Dec 74,
SYRACUSE UNIV N Y DEPT OF INDUSTRIAL ENGINEERING AND OPERATIONS RESEARCH
Pagination or Media Count:
The overall purpose of this research project is to investigate the feasibility of solving a system of linear equations of the matrix form AX B using an associative processor. The special consideration here is that the system will be very large, perhaps on the order of 2000 to 3000 unknowns. This report is concerned with the first phase of the research, that of developing timing figures for the inversion of the A matrix through Gauss Elimination using the STARAN architecture. The inversion of the full A matrix is a conservative approach but will yield upper bound results which are of importance in this feasibility study. Other methods will be considered in the future.
- Theoretical Mathematics
- Computer Programming and Software