ePubs

The open archive for STFC research publications

Full Record Details

Persistent URL http://purl.org/net/epubs/work/10917475
Record Status Checked
Record Id 10917475
Title An out-of-core sparse Cholesky solver
Contributors
Abstract Direct methods for solving large sparse linear systems of equations are popular because of their generality and robustness. Their main weakness is that the memory they require usually increases rapidly with problem size. We discuss the design and develoment of the first release of a new symmetric direct solver that aims to circumvent this limitation by allowing both the system matrix and its factors to be stored externally. The code, which is writen in Fortran and called HSL_MA77, implements a multifunctional algorithm. The first release is for positive-definite systems and performs a Cholesky factorization. Special attention is paid to the use of efficient dense linear algebra kernal codes that handle the full-matrix operations on the frontal matrix and to input/output operations. These are performed using a separate package that provides a virtual-memory system and all ows the data to be spread over many files; for very large problems these may be held on more than one devise. Numerical results are presented for a collection of 26 large real-world problems, all of which were solved successfully.
Organisation CCLRC , CSE , CSE-NAG
Keywords sparse symmetric linear systems , out-of-core-solver , multifrontal , Cholesky
Funding Information
Related Research Object(s): 37501
Licence Information:
Language English (EN)
Type Details URI(s) Local file(s) Year
Report RAL Technical Reports RAL-TR-2006-013. STFC, 2006. RAL-TR-2006-013.pdf 2006