===firstname: Ruipeng ===firstname3: Eugene ===affil6: ===lastname3: Vecharynski ===email: li50@llnl.gov ===keyword_other2: ===lastname6: ===affil5: Department of Computer Science & Engineering, University of Minnesota, Twin Cities ===lastname4: Yang ===lastname7: ===affil7: ===postal: Center for Applied Scientific Computing, Lawrence Livermore National Laboratory, P. O. Box 808, L-561, Livermore, CA 94551 ===ABSTRACT: Polynomial filtering can provide a highly effective means of computing all eigenvalues of a real symmetric (or complex Hermitian) matrix that are located in a given interval, anywhere in the spectrum. This work presents a technique for tackling this problem by combining a Thick-Restart version of the Lanczos algorithm with deflation (`locking') and a new type of polynomial filters. Thick restarting is employed to limit the cost of orthogonalization. The polynomial filter that maps eigenvalues within a given interval to eigenvalues with the largest magnitude of the transformed problem, is obtained from a least-squares approximation to an appropriately centered Dirac-$\delta$ distribution. The resulting algorithm can be utilized in a `spectrum-slicing' approach whereby a very large number of eigenvalues and associated eigenvectors of the matrix are computed by extracting eigenpairs located in different sub-intervals independently from one another. Numerical experiments show that such a construction yields effective polynomial filters which along with a Thick-Restart Lanczos procedure enable desired eigenpairs to be computed efficiently. ===affil3: Computational Research Division, Lawrence Berkeley National Laboratory ===title: A Thick-Restart Lanczos algorithm with polynomial filtering for Hermitian eigenvalue problems ===affil2: Department of Computer Science & Engineering, University of Minnesota, Twin Cities ===lastname2: Xi ===firstname4: Chao ===keyword1: Eigenvalue and singular value methods and applications ===workshop: no ===lastname: Li ===firstname5: Yousef ===keyword2: NOT_SPECIFIED ===otherauths: ===affil4: Computational Research Division, Lawrence Berkeley National Laboratory ===competition: no ===firstname7: ===firstname6: ===keyword_other1: ===lastname5: Saad ===affilother: ===firstname2: Yuanzhe