Send mail to: mgnet@cs.yale.edu for the digests mgnet-requests@cs.yale.edu for comments or help Anonymous ftp repository: casper.cs.yale.edu (128.36.12.1) World Wide Web: file://na.cs.yale.edu/pub/mgnet/www/mgnet.html or http://www.cerfacs.fr Today's editor: Craig Douglas (douglas-craig@cs.yale.edu) Volume 4, Number 11 (approximately November 30, 1994) Today's topics: Student Paper Deadline for Copper Mountain Meeting Looking for someone using MG and FEM to solve the NS eqs. WWW and Copper German Scientific Computing Home Page Multigrid and Domain Decomposition Days at CERFACS in 1995 Flame paper with detailed chemistry A new book ------------------------------------------------------- Date: Wed, 30 Nov 94 14:50:01 EST From: Craig Douglas Subject: Student Paper Deadline for Copper Mountain Meeting ----> December 1, 1994 <---- ^^^^^^^^^^^^^^^^ Send last minute papers (PostScript) by e-mail to ccmm@boulder.colorado.edu A reminder of what this: Travel and lodging assistance will be awarded to students judged to have the best research papers. Papers with original research due mainly to the student must be received before December 1, 1994. They must be singly authored and no more than 10 pages in length. Sending only an abstract is unacceptable. A panel of judges will determine the best papers. Authors of the three best papers will be given a travel stipend, free lodging, and registration, and will be expected to present their papers in a special session at the conference. Several other students who participate in the competition may be awarded free lodging and registration, depending on availability of funds. ------------------------------------------------------- Date: Fri, 4 Nov 94 12:33:53 -0800 From: gpli@jupiter.lbl.gov (guoping li) Subject: Looking for someone using MG and FEM to solve the NS eqs. Hi, I want to find someone who have simple research codes to solve the Navier-Stokes equations, based on multigrid AND finite element methods, AND will allow me to use their codes on the conditions they specify. I want to use the codes solely for my research in heat transfer. I am not look for a commercial code. Any direction and suggestion on how to find them are welcome. Thanks. Guoping ------------------------------------------------------- Date: Sat, 12 Nov 1994 11:13:32 -0700 From: stevem@boulder.colorado.edu (Steve McCormick) Subject: WWW and Copper We are now on WWW: http://amath.colorado.edu/appm/faculty/ccmm/Home.html is the WWW address for Copper info. Craig, can you (and anyone else for that matter) please put this link in whatever home page (MGNet, etc.) you would? Also, any suggestions for new/revised info or format are welcome... steve Editor's Note: I have updated the MGNet WWW entries at Yale, CERFACS, ------------- and IBM so that the postcard button of Copper Mountain goes over to Steve's URL. ------------------------------------------------------- Date: Mon, 21 Nov 1994 19:36:31 +0100 From: Ulrich Ruede Subject: German Scientific Computing Home Page We have recently started to provide a collection of WWW pages accessible via URL http://www5.informatik.tu-muenchen.de/sci-comp/home.html These documents collect information about Scientific Computing in Germany. We invite the active participation of the Scientifc Computing Community both within Germany and internationally. All feedback to scicomp@informatik.tu-muenchen.de is welcome. Please help to make this service useful by sending us the information you have! Uli Ruede ------------------------------------------------------- Date: Wed, 30 Nov 94 15:35:42 EST From: Craig Douglas Subject: Multigrid and Domain Decomposition Days at CERFACS in 1995 CERFACS is having a special year on linear algebra in 1995. Part of this will be 4 workshops on particular topics. The first of these will be April 10-13 and will focus on iterative methods. This workshop is being organized by Iain Duff, Luc Giraud, and myself. The tentative schedule is the following: April 10 an industrial tutorial, April 11 Krylov space methods, April 12 multigrid methods April 13 domain decomposition methods A collection of experts will give talks on these days. If you are interested in attending, please send e-mail to Dr. Chiara Puglisi: puglisi@cerfacs.fr ------------------------------------------------------- Date: Wed, 30 Nov 94 15:07:13 EST From: Craig Douglas Subject: Flame paper with detailed chemistry Detailed Chemistry Modeling of Laminar Diffusion Flames on Parallel Computers Craig C. Douglas IBM T. J. Watson Research Center and Department of Computer Science, Yale University Alexandre Ern Department of Mechanical Engineering, Yale University, and CERMICS Mitchell D. Smooke Department of Mechanical Engineering, Yale University Abstract: We present a numerical simulation of an axisymmetric, laminar diffusion flame with finite rate chemistry on serial and distributed memory parallel computers. We use the total mass, momentum, energy, and species conservation equations with the compressible Navier-Stokes equations written in vorticity-velocity form. The computational algorithm for solving the resulting nonlinear coupled elliptic partial differential equations involves damped Newton iterations, Krylov-type linear system solvers, and adaptive mesh refinement. The results presented here are the first in which a lifted diffusion flame structure is obtained on a nonstaggered grid. The numerical solution is in very good agreement with previous numerical and experimental data. Key words: combustion, finite rate chemistry, vorticity-velocity, nonlinear methods, iterative methods, parallel computers. AMSMOS Classification: 80A32, 80-08, 65C20, 65N20, 65F10. Editor's Note: in mgnet/papers/Douglas-Ern-Smooke/detailed.ps and ------------- mgnet/papers/Douglas-Ern-Smooke/detailed.abs. There are some color figures that have been left out of the online version. Contact one of the authors for these. ------------------------------------------------------- From: AOH Axelsson Date: Sat, 12 Nov 1994 18:09:43 GMT Subject: A new book New book available ITERATIVE SOLUTION METHODS Cambridge University Press, 1994 Author: O. Axelsson 654 pages ISBN 0-521-44524-8 hardback List price (Europe) \pounds 50. The first seven chapters and appendix A and B can be used as a textbook for a master-class course in numerical linear algebra. It presents also some basic theory in linear algebra. There is an abundance of exercises, some of which presenting additional methods in a self-programmed style. The remaining six chapters present recent results in iterative solution methods and can be used in a more advanced course. It can also be useful for research- or application-oriented students and scientists. CONTENTS 1.Direct solution methods 1.1 Introduction: Networks and structures 1.2 Gaussian Elimination and Matrix Factorization 1.3 Range and Nullspace 1.4 Practical Considerations 1.5 Solution of Tridiagonal Systems of Equations Exercises References 2. Theory of Matrix Eigenvalues 2.1 The Minimal Polynomial 2.2 Selfadjoint and Unitary Matrices 2.3 Matrix Equivalence (Similarity Transformations) 2.4 Normal and H-normal Matrices Exercises References 3. Positive Definite Matrices, Schur Complements and General Eigenvalue Problems 3.1 Positive Definite Matrices 3.2 Schur Complements 3.3 Condition Numbers 3.4 Estimates of Eigenvalues of Generalized Eigenvalue Problems 3.5 Congruent Transformations 3.6 Quasisymmetric Matrices Exercises References 4. Reducible and Irreducible Matrices and the Perron-Frobenius Theory for Nonnegative Matrices 4.1 Reducible and Irreducible Matrices 4.2 Gershgorin Type Eigenvalue Estimates 4.3 The Perron-Frobenius Theorem 4.4 Rayleigh Quotient and Numerical Range 4.5 Some Estimates of the Perron-Frobenius Root of Nonnegative Matrices 4.6 A Leontief Closed Input-Output Model Exercises References 6. M-Matrices, Convergent Splittings and the SOR Method 6.1 M-Matrices 6.2 Convergent Splittings 6.3 Comparison Theorems 6.4 Diagonally Compensated Reduction of Positive Matrix Entries 6.5 The SOR Method Exercises References 7. Incomplete Factorization Preconditioning Methods 7.1 Point Incomplete Factorization 7.2 Block Incomplete Factorization; Introduction 7.3 Block Incomplete Factorization of M-Matrices 7.4 Block Incomplete Factorization of Positive Definite Matrices 7.5 Incomplete Factorization Method for Block H-Matrices 7.6 Inverse Free Form for Block Tridiagonal Matrices 7.7 Symmetrization of Preconditioners and the SSOR and ADI Methods Exercises References 8. Approximate Matrix Inverses and Corresponding Preconditioning Methods 8.1 Two Methods of Computing Approximate Inverses of Block Bandmatrices 8.2 A Class of Methods for Computing Approximate Inverses of Matrices 8.3 A Symmetric and Positive Definite Approximate Inverse 8.4 Combinations of Explicit and Implicit Methods 8.5 Methods of Matrix Action 8.6 Decay Rates of (Block-) Entries of Inverse of (Block-) Tridiagonal s.p.d. Matrices References 9. Block Diagonal and Schur Complement Preconditionings 9.1 The C.B.S. Constant 9.2 Block-Diagonal Preconditioning 9.3 Schur Complement Preconditioning 9.4 Full Block-Matrix Factorization Methods 9.5 Indefinite Systems References 10. Estimates of Eigenvalues and Condition Numbers for Preconditioning Matrices 10.1 Upper Eigenvalue Bounds 10.2 Perturbation Methods 10.3 Lower Eigenvalue Bounds for M-Matrices 10.4 Upper and Lower Bounds of Condition Numbers 10.5 Asymptotic Estimates of Condition Numbers for Second-Order Elliptic Problems References 11. Conjugate Gradient and Lanczos-Type Methods 11.1 The Three-Term Recurrence Form of the Conjugate Gradient Method 11.2 The Standard Conjugate Gradient Method 11.3 The Lanczos Method for Generating A-Orthogonal Vectors References 12. Generalized Conjugate Gradient Methods 12.1 Generalized Conjugate Gradient, Least Squares Methods 12.2 Orthogonal Error Methods 12.3 Generalized Conjugate Gradient Methods and Variable (Nonlinear) Preconditioners References 13. The Rate of Convergence of the Conjugate Gradient Method 13.1 Rate of Convergence Estimates based on Min Max Approximations 13.2 Estimates based on the Condition Number 13.3 An Estimate Based on a Ratio Involving the Trace and the Determinant 13.4 Estimates of the Rate of Convergence Using Different Norms 13.5 Conclusions References Appendices: A Matrix Norms, Inherent Errors and Computation of Eigenvalues B Chebyshev Polynomials C Some Inequalities for Functions of Matrices Index ------------------------------ End of MGNet Digest **************************