Title :
Multicoloring for Fast Sparse Matrix-Vector Multiplication in Solving PDE Problems
Author :
Wang, H.C. ; Hwang, Kai
Author_Institution :
University of Southern California, Los Angeles, USA
Abstract :
A new multicoloring technique is proposed for parallel sparse matrix-vector multiplication, which dominates the computing cost of iterative PDE (partial differential equation) solvers. The new technique enables parallel solution of grid-structured nonsymmetric PDE problems on shared-memory multiprocessors through resolving memory access conflicts by multiple processors. The coloring scheme is formulated as an algebraic mapping which can be implemented with low overhead.
Keywords :
Concurrent computing; Costs; Gradient methods; Mesh generation; Multiprocessing systems; Packaging; Parallel processing; Partial differential equations; Sparse matrices; Testing; Parallel processing; conjugate gradient methods; matrix-vector multiplication; multicoloring; partial differential equations.;
Conference_Titel :
Parallel Processing, 1993. ICPP 1993. International Conference on
Conference_Location :
Syracuse, NY, USA
Print_ISBN :
0-8493-8983-6
DOI :
10.1109/ICPP.1993.119
