Scalable parallel sparse factorization with left-right looking strategy on shared memory multiprocessors
An efficient sparse LU factorization algorithm on popular shared memory multiprocessors is presented. Interprocess communication is critically important on these architectures—the algorithm introduces O(n) synchronization events only. No global barrier is used and a completely asynchronous scheduling scheme is one central point of the implementation. The algorithm aims at optimizing the single node performance and minimizing the communication overhead. It has been successfully tested on SUN Enterprise, DEC AlphaServer, SGI Origin 2000, Cray T90, J90, and NEC SX-4 parallel computers, delivering up to 2.3 GFlop/s on an eight processor DEC AlphaServer for medium-size semiconductor device simulations and structural engineering problems.
Unable to display preview. Download preview PDF.
- 3.D. Fokkema, Subspace methods for linear, nonlinear, and eigen problems., PhD thesis, Utrecht University, 1996.Google Scholar
- 4.Integrated Systems Engineering AG, DESSIS-iseReference Manual, ISE Integrated Systems Engineering AG, 1998.Google Scholar
- 5.Integrated Systems Engineering AG, DESS-iseReference Manual, ISE Integrated Systems Engineering AG, 1998.Google Scholar
- 6.G. Karypis and V. Kumar, Analysis of multilevel graph algorithms, Tech. Rep. MN 95–037, University of Minnesota, Department of Computer Science, Minneapolis, MN 55455, 1995.Google Scholar
- 7.A. Liegmann, Efficient Solution of Large Sparse Linear Systems, PhD thesis, ETH Zürich, 1995.Google Scholar
- 11.E. Rothberg, Exploiting the memory hierarchy in sequential and parallel sparse Cholesky factorization, PhD thesis, Stanford University, 1992. STAN-CS-92-1459.Google Scholar
- 12.Y. Saad, Iterative Methods for Sparse Linear Systems, PWS Publishing Company, 1996.Google Scholar
- 13.O. Schenk, K. Gärtner, and W. Fichtner, Efficient sparse LU factorization with left-right looking strategy on shared memory multiprocessors, Tech. Rep. 98/40, Integrated Systems Laboratory, ETH Zurich, Swiss Fed. Inst. of Technology (ETH), Zurich, Switzerland, Submitted to BIT Numerical Mathematics, 1998.Google Scholar
- 14.G. Sleijpen, H. Van Der Vorst, and D. Fokkema, BiCGSTAB(l) and other hybrid Bi-CG methods, Tech. Rep. TR Nr. 831, Department of Mathematics, University Utrecht, 1993.Google Scholar