An Implementation of Parallel Eigenvalue Computation Using Dual-Level Hybrid Parallelism

This paper describes a hybrid two-level parallel method with MPI/OpenMP for computing the eigenvalues of dense symmetric matrices on cluster of SMP's environments. The eigenvalue computation is Based on both the Householder tridiagonalization method and a divide-and-conquer algorithm of tridiagonal eigenproblem. In hybrid parallel design, We take a coarse-grain approach to OpenMP shared-memory parallelization, which keeps BLAS-3 operations in tridiagonalization. Moreover, dynamic work sharing is used in the divide-and-conquer algorithm of tridiagonal eigenproblem. So the amount of synchronization has also been reduced, and these could have an effect on the load balance. In addition, we analyze the communication overhead between hybrid MPI/ OpenMP and pure MPI. An experimental analysis on the Deepcomp6800 shows the hybrid algorithm performs good scalability.

[1]  R. J. Littlefield,et al.  Performance of a fully parallel dense real symmetric eigensolver in quantum chemistry applications , 1995 .

[2]  R.D. Loft,et al.  Terascale Spectral Element Dynamical Core for Atmospheric General Circulation Models , 2001, ACM/IEEE SC 2001 Conference (SC'01).

[3]  Gerhard Wellein,et al.  Communication and Optimization Aspects of Parallel Programming Models on Hybrid Architectures , 2003, Int. J. High Perform. Comput. Appl..

[4]  J. Cuppen.  A divide and conquer method for the symmetric tridiagonal eigenproblem , 1980 .

[5]  P. Alpatov,et al.  PLAPACK Parallel Linear Algebra Package Design Overview , 1997, ACM/IEEE SC 1997 Conference (SC'97).

[6]  B. Parlett,et al.  Multiple representations to compute orthogonal eigenvectors of symmetric tridiagonal matrices , 2004 .

[7]  Xiaobai Sun,et al.  The PRISM project: infrastructure and algorithms for parallel eigensolvers , 1993, Proceedings of Scalable Parallel Libraries Conference.

[8]  Franck Cappello,et al.  MPI versus MPI+OpenMP on the IBM SP for the NAS Benchmarks , 2000, ACM/IEEE SC 2000 Conference (SC'00).

[9]  Stanley C. Eisenstat,et al.  A Divide-and-Conquer Algorithm for the Symmetric Tridiagonal Eigenproblem , 1995, SIAM J. Matrix Anal. Appl..

[10]  James Demmel,et al.  Applied Numerical Linear Algebra , 1997 .

[11]  Elizabeth R. Jessup,et al.  Toward an Efficient Parallel Eigensolver for Dense Symmetric Matrices , 1998, SIAM J. Sci. Comput..

[12]  Jack Dongarra,et al.  ScaLAPACK Users' Guide , 1987 .

[13]  David J. Kuck,et al.  A Parallel QR Algorithm for Symmetric Tridiagonal Matrices , 1977, IEEE Transactions on Computers.

[14]  Philip K. McKinley,et al.  A Scalable Eigenvalue Solver for Symmetric Tridiagonal Matrices , 1995, Parallel Comput..

[15]  Xuebin Chi,et al.  Solving the Symmetric Tridiagonal Eigenproblem Using MPI/OpenMP Hybrid Parallelization , 2005, APPT.

[16]  Zhonggang Zeng,et al.  The Laguerre Iteration in Solving the Symmetric Tridiagonal Eigenproblem, Revisited , 1994, SIAM J. Sci. Comput..

[17]  Zeki Demirbilek,et al.  Dual-Level Parallel Analysis of Harbor Wave Response Using MPI and OpenMP , 2000, Int. J. High Perform. Comput. Appl..

[18]  Suchuan Dong,et al.  Dual-level parallelism for high-order CFD methods , 2004, Parallel Comput..

[19]  Le N. Ly,et al.  Coastal Ocean Modeling of the U.S. West Coast with Multiblock Grid and Dual-Level Parallelism , 2001, ACM/IEEE SC 2001 Conference (SC'01).

[20]  Liu Wei.  The concept of node-oriented speedup on SMP cluster , 2000 .