Subject: NA Digest, V. 93, # 16 NA Digest Sunday, April 18, 1993 Volume 93 : Issue 16 Today's Editor: Cleve Moler The MathWorks, Inc. moler@mathworks.com Today's Topics: References for SIMD Linear Algebra Seeking a Better Expansion Looking for the Second Edition of Lawson and Hanson A Great Bargain C++ for Numerical Analysis Hellenic Mathematical Society, 75 Years 1993 PVM Users' Group Meeting Numerical Methods for the Solution of the Maxwell Equations ACM-SIAM Symposium on Discrete Algorithms Position at CERCA, Montreal Contents, SIAM Math Analysis Contents, Acta Numerica Submissions for NA Digest: Mail to na.digest@na-net.ornl.gov. Information about NA-NET: Mail to na.help@na-net.ornl.gov. ------------------------------------------------------- From: Glenn R. Luecke Date: Mon, 12 Apr 93 08:44:29 CDT Subject: References for SIMD Linear Algebra I am looking for references for papers dealing with parallel linear algebra routines (especially for nonsparse LU factorization, Cholesky factorization, SVD and eigensystem solvers) for SIMD computers, such as Thinking Machines' CM-2 and MasPar's MP-1/MP-2 computers. So far I have not found any such references and would greatly appreciate it if anyone could provide me with such. Thanks! Glenn Luecke GM.GRL@ISUMVS.IASTATE.EDU 291 Durham Center, Iowa State University, Ames, IA 50011 ------------------------------ From: Stewart Levin Date: Mon, 12 Apr 93 13:08:52 CDT Subject: Seeking a Better Expansion A problem arising in exploration geophysics is that of dereverberation. We probe an earth structure X with an impulsive signal and receive a response U=X(1+X)^-1 due to (free surface) reverberation. One way to recover X from U is to form U(1-U)^-1 = U + U^2 + U^3 + U^4 + ... which in the limit recovers X. A problem that arises in practice is that (1-U)^-1 is not practical to compute directly and so we employ the MacLaurin expansion instead. Taking partial sums of the MacLaurin series has a bad side effect of boosting later reverbera- tions as it suppresses earlier ones, e.g. partial sums look like: S1 = X - X^2 + X^3 - X^4 + X^5 - X^6 + X^7 - X^8 + ... S2 = X - X^3 + 2X^4 - 3X^5 + 4X^6 - 5X^7 + 6X^8 + ... S3 = X - X^4 + 3X^5 - 6X^6 +10X^7 -15X^8 + ... S4 = X - X^5 + 4X^6 -10X^7 +20X^8 + ... S5 = X - X^6 + 5X^7 -15X^8 + ... S6 = X - X^7 + 6X^8 + ... S7 = X - X^8 + ... S8 = X + ... Can anyone come up with an alternative expansion, e.g. in terms of polynomial families or product representations, which has the desirable property that the coefficients of the Taylor expansion of the partial sums (or products) is always less than 1 in magnitude? -- stew@hanauma.stanford.edu -- salevin@dal.mobil.com ------------------------------ From: Chuck Lawson Date: Tue, 13 Apr 93 9:22:29 PDT Subject: Looking for the Second Edition of Lawson and Hanson To: readers of the na.digest -- I would like to buy a copy of the Second Edition of "Solving Least Squares Problems" by Lawson and Hanson, published by Prentice-Hall. It is now out of print. Please contact me if you have a copy you are willing to sell. -- Charles Lawson, clawson@math.jpl.nasa.gov ------------------------------ From: Gene Golub Date: Tue, 13 Apr 93 21:58:27 PDT Subject: A Great Bargain Dover Publications has just re-issued Herb Keller's "Numerical Methods for Two-Point Boundary Value Problems" The cost is $US 9.95. (That's not a typographical error!) The volume also includes Keller's monograph published by SIAM and several of Keller's paper. Coming soon: Isaacson and Keller. Thanks, Dover. You are doing a terrific job for our community. Gene Golub ------------------------------ From: Walter Figueiredo Mascarenhas Date: Fri, 16 Apr 93 16:20:03 BSC Subject: C++ for Numerical Analysis I would like to hear the impressions from people using C++ for Numerical Analysis. Is it true that C++'s (and C's) numerics are not as good/accurate/stable as FORTRAN's? How good are the libraries Linpack.h++, from Rogue Wave, and M++, from Dyad? Do you know other (supported and well documented) libraries? If you think that FORTRAN is THE language for numerical analysis and would not waste your time with C++, please let me know why. Thanks for your help, Walter Mascarenhas walter@ime.unicamp.br ------------------------------ From: Hellenic Math Society Date: Wed, 14 APR 93 16:44 GMT Subject: Hellenic Mathematical Society, 75 Years SCIENTIFIC EVENTS OF THE HELLENIC MATHEMATICAL SOCIETY FOR THE COMPLETION OF 75-YEARS FROM ITS ESTABLISHMENT The Hellenic Mathematical Society (HMS) on the occassion of the completion of 75-years of active multi-disciplinary scientific contribution in the Hellenic and International Mathematical Community, is organizing a series of scientific events during the year 1993. These celebratory scientific activities of HMS include: (I) MATHEMATICAL ONE-DAY MEETINGS (HMS-75-MH) A series of scientific meetings and discusions with general and specific subjects of Mathematics and its applications. These scientific (one-day) meetings will take place in Athens at regular time-periods during the year 1993. The three distinguished speakers in such a meeting will be selected from the following mathematical categories: (i) Members of the Academy of Athens and ex-Presidents of the HMS. Subject of their talks: the contribution of Greek Mathematicians in the development of the modern Mathematical Science, the activities of the HMS etc. (ii) Heads of the Departments of Mathematics of the Hellenic Universities and Technical Universities as well as distinguished Researchers-Professors of Academic Institutions. Subject of their talks: Modern programs of mathematical studies and their applications, present research activities of Departments of Mathematics, participation in Hellenic and International Research Programmes etc. (iii) Presidents and national representatives of the Mathematical Societies of the following countries: United States of America, C.I.S. (Academy of Science, Moscow), Japan, China, Britain, Germany, France, as well as national representatives from other European countries. Subject of their talks: present activities of the Scientific Societies of their countries, participation in international research programmes, development of international scientific relationships, personal research interests etc (II) HMS SCIENTIFIC RESEARCH AWARD (HMS-75-D ) The HMS announces a scientific research award for the most outstanding Doctoral Thesis in Mathematics and its applications written by Greek scientist in the period 1990-1993. This scientific distinction will be awarded to the eminent Greek Ph.D. author, whose the doctoral programme has been completed in a Hellenic University (or corresponding Academic Institution abroad) in a special ceremony of the HMS near the end of 1993, after the final decision of the HMS-75-D Evaluation Committee. The prospective candidates for the above HMS Scientific research award should sent three copies of their Doctoral Thesis (Ph.D), written in Greek or English language, and a short CV in the following address: Hellenic Mathematical Society HMS-75-D Evaluation Committee 34 Panepistimiou Street GR- 106 79, Athens G r e e c e The deadline for the submission of entries is July 31st, 1993. (III) SPECIAL VOLUME FOR THE 75-YEARS of HMS (EME-75-EE) In the framework of its foundation purposes and with the occassion of the completion of 75-years, the HMS will publish a collective scientific volume with main subject: "The contribution of the ancient Greek Mathematicians in the development of the modern Mathematical Science and its applications " In this work have been invited to participate with their distinguished contributions members of the National Mathematical Committee (Academy of Athens), University Professors and eminent Research Workers. Those who would like to submit suitable work for inclusion in this HMS-75 volume please write to HMS address given above. Professor Elias A. Lipitakis President of Hellenic Mathematical Society ------------------------------ From: Jack Dongarra Date: Wed, 14 Apr 93 12:56:57 -0400 Subject: 1993 PVM Users' Group Meeting 1993 PVM Users Group Meeting Knoxville, Tennessee May 10 - 11, Knoxville Hilton, Knoxville, TN The first PVM Users' Group Meeting will be held May 10th and 11th, 1993, at the Downtown Knoxville Hilton, Knoxville, TN. The meeting will provide a chance for users of PVM to meet each other and members of The PVM Team. The conference will include: * Invited talks by experts in fields of interest to PVM users. * Contributed papers from the PVM user community. * Technical presentations by the developers of PVM. * Social activities. Call For Papers/Posters We welcome abstracts of papers and posters for possible presentation at the Users' Group meeting. Abstracts should be approximately 300 words in length. The deadline for submitting abstracts is April 23, 1993. Abstracts will be reviewed and authors notified of their acceptance by April 30. Please submit abstracts to: Jan Jones Computer Science Department University of Tennessee Knoxville, TN 37996 phone: 615-974-8295, fax: 615-974-8296 or to: pvmug@cs.utk.edu Additional information on the Users' Group meting can be found in netlib. Send email to netlib@ornl.gov and in the mail message type: send pvmug93 from pvm ------------------------------ From: Rainald Lohner Date: Thu, 15 Apr 93 11:43:25 EDT Subject: Numerical Methods for the Solution of the Maxwell Equations SECOND INTERNATIONAL CONFERENCE AND WORKSHOP ON APPROXIMATIONS AND NUMERICAL METHODS FOR THE SOLUTION OF THE MAXWELL EQUATIONS Key Bridge Marriott Washington, DC October 25-27, 1993 CALL FOR PAPERS The organizing committee is pleased to announce the SECOND INTERNATIONAL CONFERENCE AND WORKSHOP ON APPROXIMATIONS AND NUMERICAL METHODS FOR THE SOLUTION OF THE MAXWELL EQUATIONS. The aim of the Conference is to review progress and assess new techniques for the numerical solution of the Maxwell equations, with particular emphasis on the applicability for Radar Cross Sections (RCS) calculations of perfectly and partially conducting objects. The Conference will feature invited speakers as well as contributed papers. The aim of the Workshop is to compare the different numerical methodologies currently employed for RCS calculations. Comparisons in terms of accuracy and efficiency will be conducted on a set of 24 selected test cases. Participation to the Workshop is open to anybody willing to compute and document at least two of the mandatory test cases which consist of different geometries. The Conference and Workshop will be held in Washington, DC, with the help of CMEE/SEAS-GWU (US) and GAMNI-SMAI (Europe). Organizing Committee: R. Glowinski (Univ. of Houston), R. Lohner (The George Washington University, Co-Chair), J. Periaux (GAMNI-SMAI, Co-Chair), V. Shankar (Rockwell Intl. Science Center) Scientific Committee: R. Agarwal (McDonnell Douglas); J. Ambrosiano (LLNL); G. Borghiotti (GWU); J. Crowley (DARPA); A. Drobot (SAIC, McLean); J. Fratemico (SAIC, San Diego); P. Goorjian (NASA Ames); M. Hafez (UC Davies); P. Joly (INRIA); J. Kong (MIT); B. Mercier (Aerospatiale);R. Mittra (U. Illinois); K. Morgan (Swansea); J.C. Nedelec (Ecole Polytechnique); J.P. Priou (MOD, France); V. Rokhlin (Yale); T.K. Sarkar (Syracuse); B. Scheurer (CEA); J. Shang (WPAFB); J. Ziolkowski (U. Arizona) ABSTRACT DEADLINE : May 1st, 1993. Proposed title and 2-3 pages extended abstract sould be sent to: US and non-European participants: European participants: R. Lohner J. Periaux Office of Conferences and Institutes GAMNI The George Washington University INRIA, Projet Menusi 2003 G Street, NW Domaine de Voluceau Washington, DC 20052 Rocquencourt B.P. 105 78153 Le Chesnay, Cede FRANCE FAX: 202-994-7400 FAX: 33-1-39.63.53.3 TEL: 202-994-7238 TEL: 33-1-39.63.54.82 e-mail: lohner@seas.gwu.edu e-mail: periaux@menusin.inria.fr WORKSHOP TEST CASES Instructions for retrieving the test-cases data: Through anonymous ftp: On your home machine "cd" to the directory where you want the data to go; then enter the following: ftp ftp.gwu.edu at the prompt, enter: anonymous and then enter your e-mail address as the password. Then type: cd pub/emworkshop binary prompt mget * quit That should do it. ------------------------------ From: Trini Flores Date: Fri, 16 Apr 93 08:44:51 EST Subject: ACM-SIAM Symposium on Discrete Algorithms SODA 94 - - CALL FOR PAPERS Fifth Annual ACM-SIAM Symposium on Discrete Algorithms Sponsored by ACM-SIGACT and SIAM Activity Group on Discrete Mathematics January 23-25, 1994 Key Bridge Marriott Hotel Arlington, Virginia Organizer and Chair: Daniel D. Sleator, Carnegie-Mellon University The symposium is for those interested in the use, design, and analysis of algorithms. Papers about algorithms (sequential, parallel, distributed, randomized, etc.) in all areas of application are invited. Papers will be selected for presentation based on extended abstracts. Authors wishing to submit a paper should send twelve (12) copies of an extended abstract (not a full paper) by Tuesday, July 6, 1993 to: SIAM Conference Coordinator 3600 University City Science Center Philadelphia, PA 19104-2688 USA Abstracts should begin with a succinct statement of the problems that are considered in the paper, the main results achieved, an explanation of the significance of the work, and a comparison to past research. This material should be easily understood by nonspecialists. Technical development, directed toward the specialist, should follow as appropriate. The entire extended abstract must not exceed ten double-spaced typed pages (about 12,000 bytes). Abstracts that deviate significantly from these guidelines risk rejection without consideration of their merits. Abstracts must be received by the July 6 deadline otherwise submissions will not be considered. For additional information, contact SIAM. Telephone: 215-382-9800 Fax: 215-386-7999 E-Mail: meetings@siam.org ------------------------------ From: Maurice.Meneguzzi@cerfacs.fr Date: Thu, 15 Apr 1993 16:10:29 +0200 Subject: Position at CERCA, Montreal CERCA 5160, boul. Decarie, bureau 400 CEntre de Recherche Montreal (Quebec) Canada H3X 2H9 en Calcul Applique Telephone:(514) 369-5224 Fax:(514) 369-3880 CEntre for Research on Computation and its Applications JOB ANNOUNCEMENT: RESEARCH PROFESSIONAL FOR LARGE SCALE PROBLEM SOLVING CERCA invites applications for the position of "Research Professional" in the area of large scale problem solving. Candidates for the position, who must have a Master's or a doctoral degree, are required to have training in the area of numerical methods and, in particular, to be familiar with iterative methods, preconditioning techniques and their application to the finite element method. Experience with parallel computing and vectorization would be an asset. The selected candidate will be expected to contribute to the refinement of state-of-the-art techniques for the solution of large scale problems associated with numerical simulation in fluid mechanics. He will also be expected to cooperate in the operation of applied projects and in the development of new methods. CERCA is a nonprofit corporation. It was established in January 1992, following an initiative from E'cole Polytechnique, Concordia University, McGill University, and the Universite' de Montre'al, with an infrastructure grant from the Ministe`re de l'Enseignement supe'rieur et de la Science du Quebec. Currently, Alcan International, Bombardier-Canadair, CANMET (Energy, Mines and Resources Canada), Environment Canada (AES, Dorval), GE Canada, GEC Alsthom E'lectrome'canique, Hydro-Que'bec (Equipment Group), and Pratt & Whitney Canada are regular industrial members of CERCA. In addition, Spar Aerospace and Bio-Mega are associate industrial members. Interested candidates should submit a detailed re'sume', and names, addresses and phone numbers of three professional references, to Professor Andre' Biron, Associate Director, CERCA, 5160 boul. De'carie, Suite 400, Montre'al, Que'bec H3X 2H9, Canada. Applications received before May 14, 1993 will be assured consideration. CERCA offers competitive salaries and fringe benefits to its employees, and an excellent environment for professional growth. In accordance with Canadian immigration requirements, priority will be given to citizens and permanent residents of Canada. This advertisement applies equally to women and men. ------------------------------ From: SIAM Date: Wed, 14 Apr 93 16:51:23 EST Subject: Contents, SIAM Math Analysis Contents, SIAM J. Math. Anal. July 1993 Convexity of Solutions to Some Elliptic Partial Differential Equations Antonio Greco and Giovanni Porru One-Phase Riemann Problem and Wave Interactions in Systems of Conservation Laws of Mixed Type Haitao Fan Solitary Waves in a Two-Layer Fluid with Surface Tension S.M. Sun and M. C. Shen The Linear Shallow Water Theory: A Mathematical Justification James A. Donaldson and Daniel A. Williams Nonlinear Stability and Asymptotic Behavior of Shearing Motions of a Non- Newtonian Fluid John A. Nohel and Robert L. Pego Antiplane Shearing Motions of a Visco-Plastic Solid J.M. Greenberg and Anne Nouri A Geometric Singular Perturbation Analysis of Detonation and Deflagration Waves I. Gasser and P. Szmolyan Travelling Waves for Mutualist Species Konstantin Mischaikow and Vivian Hutson Stability of Equilibria for a Class of Time-Reversible, Dn x 0(2)-Symmetric Homogeneous Vector Fields I-Heng McComb and Chjan Q. Lim Unconditional Bases of Wavelets for Sobolev Spaces Gustaf Gripenberg Least Squares Approximation by Radial Functions E. Quak, N. Sivakumar, and J. D. Ward Superharmonic Solutions for Some Second-Order Differential Equations with Singularities Alessandro Fonda, Raul Manasevich, and Fabio Zanolin Estimates for Large Deviations in Random Trigonometric Polynomials George Benke and W.J. Hendricks Selberg Integrals and Hypergeometric Functions Associated with Jack Polynomials Jyoichi Kaneko ------------------------------ From: Arieh Iserles Date: Thu, 15 Apr 93 11:58:44 BST Subject: Contents, Acta Numerica Contents of Acta Numerica 1993: Gene Allgower and Kurt Georg (Colorado State University): "Continuation and path-following" ................................1 Carl de Boor (University of Wisconsin): "Multivariate splines" ..65 Jim Demmel (University of California at Berkeley), Mike Heath (University of Illinois) and Henk van der Vorst (University of Utrecht): "Parallel numerical linear algebra" .................111 J.L.M. van Dorsselear, J.F.B.M. Kraaijevanger and M.N. Spijker (Rijksuniversiteit Leiden): "Linear stability analysis in the numerical solution of initial value problems" .................199 Michel Fortin (Universite Laval): "Finite element solution of the Navier--Stokes equations" ......................................239 Harry Yserentant (University of Tuebingen): "Old and new convergence proofs for multigrid methods" ..................................285 ------------------------------ End of NA Digest ************************** -------