Towards an Accurate Performance Modeling of Parallel SparseFactorization
Journal Article
·
· Applicable Algebra in Engineering, Communication, andComputing
OSTI ID:923277
We present a performance model to analyze a parallel sparseLU factorization algorithm on modern cached-based, high-end parallelarchitectures. Our model characterizes the algorithmic behavior bytakingaccount the underlying processor speed, memory system performance, aswell as the interconnect speed. The model is validated using theSuperLU_DIST linear system solver, the sparse matrices from realapplications, and an IBM POWER3 parallel machine. Our modelingmethodology can be easily adapted to study performance of other types ofsparse factorizations, such as Cholesky or QR.
- Research Organization:
- COLLABORATION - Parc Club OrsayU./France
- DOE Contract Number:
- DE-AC02-05CH11231
- OSTI ID:
- 923277
- Report Number(s):
- LBNL-60384; R&D Project: KS1210; BnR: KJ0101010; TRN: US200804%%1039
- Journal Information:
- Applicable Algebra in Engineering, Communication, andComputing, Vol. 18, Issue 3; Related Information: Journal Publication Date: 05/2007
- Country of Publication:
- United States
- Language:
- English
Similar Records
Data Locality Enhancement of Dynamic Simulations for Exascale Computing (Final Report)
Parallel sparse matrix computations: Wavefront minimization of sparse matrices. Final report for the period ending June 14, 1998
Parallel sparse Cholesky factorization algorithms for shared-memory multiprocessor systems
Technical Report
·
Fri Nov 29 00:00:00 EST 2019
·
OSTI ID:923277
Parallel sparse matrix computations: Wavefront minimization of sparse matrices. Final report for the period ending June 14, 1998
Technical Report
·
Mon Feb 01 00:00:00 EST 1999
·
OSTI ID:923277
Parallel sparse Cholesky factorization algorithms for shared-memory multiprocessor systems
Conference
·
Wed Jul 01 00:00:00 EDT 1992
·
OSTI ID:923277
+1 more