## NA Digest Sunday, December 10, 2000 Volume 00 : Issue 50

Today's Editor:
Cleve Moler
The MathWorks, Inc.
moler@mathworks.com

### Submissions for NA Digest:

Mail to na.digest@na-net.ornl.gov.

Mail to na.help@na-net.ornl.gov.

URL for the World Wide Web: http://www.netlib.org/na-net/na_home.html
-------------------------------------------------------

From: Tim Davis <davis@cise.ufl.edu>
Date: Tue, 05 Dec 2000 10:53:38 -0500
Subject: Re: Flop Count Missing in Matlab R12

To count flops, we need to first know what they are. What is a flop?

LAPACK is not the only place where the question "what is a flop?" is relevant.
Sparse matrix codes are another. Multifrontal and supernodal factorization
algorithms store L and U (and intermediate submatrices, for the multifrontal
method) as a set of dense submatrices. It's more efficient that way, since the
dense BLAS can be used within the dense submatrices. It is often better
explicitly store some of the numerical zeros, so that one ends up with fewer
frontal matrices or supernodes.

So what happens when I compute zero times zero plus zero? Is that a flop
(or two flops)? I computed it, so one could argue that it counts. But it
was useless, so one could argue that it shouldn't count. Computing it allowed
me to use more BLAS-3, so I get a faster algorithm that happens to do some
useless flops. How do I compare the "mflop rate" of two algorithms that
make different decisions on what flops to perform and which of those to
include in the "flop count"?

A somewhat better measure would be to compare the two algorithms based an
external count. For example, the "true" flop counts for sparse LU
factorization can be computed in Matlab from the pattern of L and U as:

[L,U,P] = lu (A) ;
Lnz = full (sum (spones (L))) - 1 ; % off diagonal nz in cols of L
Unz = full (sum (spones (U')))' - 1 ; % off diagonal nz in rows of U
flops = 2*Lnz*Unz + sum (Lnz) ;

The same can be done on the LU factors found by any other factorization code.
This does count a few spurious flops, namely the computation a_ij + l_ik*u_kj
is always counted as two flops, even if a_ij is initially zero.

However, even with this "better" measure, the algorithm that does more
flops can be much faster. You're better off picking the algorithm with
the smallest memory space requirements (which is not always the smallest
nnz (L+U)) and/or fastest run time.

So my vote is to either leave out the the flop count, or at most return a
reasonable agreed-upon estimate (like the "true flop count" for LU, above)
that is somewhat independent of algorithmic details. Matrix multiply, for
example, should report 2*n^3, as Cleve states in his Winter 2000 newsletter,
even though "better" methods with fewer flops (Strassen's method) are
available.

Tim Davis
University of Florida
davis@cise.ufl.edu

------------------------------

From: Jack Dongarra <dongarra@cs.utk.edu>
Date: Tue, 05 Dec 2000 15:29:45 -0500
Subject: Re: Flop Count Missing in Matlab R12

This is to clarify some points in the discussion last week by Steve
Vavasis and Cleve Moler in the na-digest about counting floating point
operations. We have a project at the University of Tennessee called PAPI
(http://icl.cs.utk.edu/papi/). The Performance API (PAPI) project
specifies a standard application programming interface (API) for
accessing hardware performance counters available on most modern
microprocessors.
For years collecting performance data on applications programs has been
an imprecise art. The user has had to rely on timers with poor
resolution or granularity, imprecise empirical information on the number
of operations performed in the program in question, vague information on
the effects of the memory hierarchy, etc. Today hardware counters exist
on every major processor platform. These counters can provide
application developers valuable information about the performance of
critical parts of the application and point to ways for improving
performance. The current problem facing users and tool developers is
unavailable to the user level program. The focus of the PAPI project is
to provide an easy to use, common set of interfaces that will gain
thereby providing application developers the information they need to
tune their software on different platforms. The goal is to make it easy
for users to gain access to the counters to aid in performance analysis,
modeling, and tuning.
For more details on PAPI see
http://www.netlib.org/utk/people/JackDongarra/PAPERS/papi-sc2000.pdf

Jack

------------------------------

From: Mark Friedman <friedman@ultra.math.uah.edu>
Date: Wed, 06 Dec 2000 22:26:39 -0600
Subject: Reordering the Schur Form in MATLAB

Hi,

I am looking for a MATLAB code for reordering the Real Schur Form X=U*T*U'

In LAPACK this accomplished by routines TREXC and TRSEN.

I contacted MathWorks. They suggested me to write my own
MEX-files for LAPACK functions.

Does anyboody have these MEX-files or knows where to get them?

Mark Friedman.
friedman@math.uah.edu

P.S.

Here
> [U,T] = SCHUR(X) produces a Schur matrix T and a unitary
> matrix U so that X = U*T*U' and U'*U = EYE(SIZE(U)).
> X must be square.
> In my case X is real, so the Real Schur Form is returned.

------------------------------

From: Tom Wright <thomas.wright@linacre.oxford.ac.uk>
Date: Fri, 08 Dec 2000 18:41:34 +0000
Subject: MATLAB Pseudospectra GUI

The MATLAB Pseudospectra GUI is intended to make computing pseudospectra
fast and interactive. Some of the features include:

* Speed: once the GUI has been initialised, all computation is O(N^2)
* Interactive zooming with the mouse
* `Smart' default contour levels, and the ability to change the levels
plotted without re-computation
* Easy control over axes and grid size
* Choice of colour or black and white
* Ability to bypass the GUI and use the underlying routines as an
engine for pseudospectra computation
* Option of projecting onto a lower-dimensional subspace for extra speed

Version 1.1 is a significant update, which includes the following new
features (most importantly the link with eigs/ARPACK for computing
approximations to the pseudospectra of large sparse matrices, as seen at
Copper Mountain in April 2000):

* Connection to eigs for pseudospectra of large matrices (MATLAB 6 required)
* Direct pseudospectra computation for sparse matrices
* Improved implementation; up to 10% speedup compared to version 1.0
* Computation of eigenvalue condition numbers
* Scope for creating .mat files which can be used to easily
re-create plots for printing
* Many minor bug fixes and enhancements

http://web.comlab.ox.ac.uk/oucl/work/tom.wright/psgui/index.html

Tom Wright
Numerical Analysis Group
Oxford University

------------------------------

From: Yang Xiaoqi <mayangxq@polyu.edu.hk>
Date: Mon, 04 Dec 2000 11:10:06 +0800
Subject: Workshop in Italy on Optimization and Control

First Announcement for the International Workshop on
Optimization and Control with Applications

Erice, Italy, July 9-17, 2001
International School of Mathematics ``G. Stampacchia''

OBJECTIVES

The workshop aims to bring together eminent researchers and practitioners
working in optimization, optimal control and their applications for exchanging
information and ideas on the latest development in these fields.

Topics include, but are not limited to,

optimal control
unconstrained optimization
constrained optimization
complementarity and variational inequalities
equilibrium problems
semi-definite programs
semi-infinite programs
matrix functions and equations
nonsmooth optimization
generalized convexity and generalized monotonicity
and their applications.

LECTURES: The workshop consists of invited lectures
(1 hour) and contributed lectures (1/2 hour). Proceedings
including the invited lectures and a selection of contributed
lectures will be published.

INVITED LECTURES include:
Alfred Auslender, Universit\'e Claude Bernard Lyon1, France
James Burke, University of Washington, USA
Guang-Ya Chen, Chinese Academy of Sciences, China
Christian Kanzow, University of Hamburg, Germany
Jose Mario Mart\'inez, University of Campinas, Brazil
Antonino Maugeri, Universit\`a di Catania, Italy
Jong-Shi Pang, The Johns Hopkins University, USA
Elijah Polak, University of California at Berkeley, USA
Tamas Rapcsak, Hungarian Academy of Sciences, Hungary
Kees Roos, Technology University of Delft, The Netherlands
Siegfried Schaible, University of California at Riverside, USA
Klaus Schittkowski, University of Bayreuth, Germany
Alexander Shapiro, Georgia Institute of Technology, USA
Emilio Spedicato, University of Bergamo, Italy
Jie Sun, National University of Singapore, Singapore
Michal Todd, Cornell University, USA
Paul Tseng, University of Washington, USA
Xiang-Sun Zhang, Academy of Mathematics and System Sciences, China

DIRECTOR OF THE SCHOOL: Franco Giannessi (Universit\`a di Pisa)

DIRECTORS OF THE WORKSHOP:
Liqun Qi (The Hong Kong Polytechnic University),
Kok Lay Teo (The Hong Kong Polytechnic University),
E-mail: maqilq@polyu.edu.hk and mateokl@polyu.edu.hk

HOW TO PARTICIPATE: There will be several invited sessions,
and a limited number of places available for contributed
participants. Those who wish to attend the workshop
should write as early as possible to:

Ms. Eva Yiu
Department of Applied Mathematics
The Hong Kong Polytechnic University
Hung Hom, Kowloon
Hong Kong
E-mail: maevayiu@polyu.edu.hk

FURTHER INFORMATION: e-mail: Eva Yiu (maevayiu@polyu.edu.hk)
webpage:
http://www.polyu.edu.hk/~ama/events/conference/EriceItaly-OCA2001/an1.html
or contact one of the directors of the workshop.

------------------------------

From: Yves Robert <Yves.Robert@ens-lyon.fr>
Date: Mon, 04 Dec 2000 10:19:06 +0100
Subject: Scheduling and Load Balancing Workshop

Call for Papers, Euro-Par 2001
Manchester, UK, August 28 - 31, 2001
http://www.man.ac.uk/europar/

Topic 03: SCHEDULING AND LOAD BALANCING

Scheduling and load balancing techniques are key areas in the quest
for performance in parallel and distributed applications. Such
techniques can be provided either at the application level, or at the
system level, and both scenarios are of interest for this workshop.

In application level scheduling and load balancing, the mapping of a
parallel computation onto a parallel computer system is one of the
most important issues in the design of efficient parallel algorithms.
Of particular interest are irregular data structures, where the
problem of distributing the workload evenly onto parallel computing
systems can be complex.

Another subject of special relevance concerns the development of
particular characteristics of the underlying parallel computer,
facilitating the development of portable applications.

Support at the system level for scheduling and load balancing is also
of interest. Areas include experiences with checkpoint/migration
facilities, load estimation, and support for special message passing
libraries like PVM or MPI.

Networks of Workstations (NOWs) are becoming increasingly popular
platforms for parallel computing. In contrast to traditional parallel
computing, the situation in workstation clusters is characterized by
heterogeneity, the presence of interactive users, and a higher failure
rate. This has to be taken into account when developing scheduling and
must often be employed. Furthermore, the advent of the Computational
Grid makes it possible to use combinations of NOWs, parallel computers,
and storage devices distributed over a wide area. This type of
environment leads to new scheduling problems (e.g. distributed
databases, mixed batch/interactive resources, etc.).

Finally, theoretical results which can be used as solid foundations
for designing efficient and robust scheduling algorithms are
particularly welcome.

FOCUS

Theoretical foundations of scheduling algorithms
Application-specific graph partitioning
Parallel graph partitioning algorithms
Scheduling algorithms
Tools and environments for load balancing and scheduling
Scheduling on heterogeneous nodes (e.g. with different amounts of
memory or processor speeds)
Load balancing and middleware (CORBA, DCE, Java, ...)

TOPIC COMMITTEE:

Global chair: Yves Robert
Laboratoire LIP, UMR CNRS-INRIA 5668
Ecole Normale Superieure de Lyon, France
E-mail: Yves.Robert@ens-lyon.fr
URL: http://www.ens-lyon.fr/~yrobert

Multimedia Technology Research Center
Department of Computer Science
The Hong Kong University of Science and Technology, Hong Kong

Vice-chair: : Henri Casanova
Computer Science Department
University of California at San Diego (UCSD), USA
E-mail: casanova@cs.ucsd.edu
URL: http://www-cse.ucsd.edu/~casanova

Local chair: Rupert Ford
Centre for Novel Computing,
Department of Computer Science,
The University of Manchester, UK
E-mail: rupert@cs.man.ac.uk
URL: http://www.cs.man.ac.uk/cnc

------------------------------

From: Andre Weideman <weideman@dip.sun.ac.za>
Date: Mon, 4 Dec 2000 12:37:45 +0200 (GMT)
Subject: Symposium in South Africa on Numerical and Applied Mathematics

Second Announcement of SANUM2001

This is a call for papers for the 25th annual South African
Symposium on Numerical and Applied Mathematics that will be
held in Stellenbosch, South Africa, April 9-11, 2001.

Numerical analysts, applied mathematicians and engineers
are invited to attend and present twenty minute talks
on their work. All contributions in the areas of computational,
applied, or engineering mathematics will be suitable.

We are happy to announce that Greg Baker (Ohio State),
John Burns (Virginia Tech), Kathy Driver (Wits),
Nick Higham (Manchester), and Belinda King (Virginia Tech)
will present lectures, while we are awaiting confirmation
from several other well-known local and international speakers.

Details related to registration, accommodation, transport, and
tourist attractions may be found on the recently updated web page

http://dip.sun.ac.za/sanum2001

E-mail can also be sent to members of the organizing committee:

Ben Herbst (herbst@ibis.sun.ac.za)
Andre Weideman (weideman@ing.sun.ac.za)
Karin Goosen (karin@goose.sun.ac.za)

------------------------------

From: Robert B. Silver <rsilver@mbl.edu>
Date: Tue, 05 Dec 2000 14:41:07 -0500
Subject: Modeling of Biological Systems Course at Woods Hole

Good afternoon.

The use of modeling in biological and medical sciences is increasing at a
quickening pace. Modeling is opening new perspectives and appreciation to
biological entities and processes that are often not accessible by
available "wet" methods or in an acceptable time frame.

Areas being advanced by modeling include: molecular structure and dynamics,
cell structure and dynamics, genomics, proteomics, cellomics, metabolic
networks, biophysics of ion channels, physiology and epidemiology,
population biology and ecology.

Recently, we launched a multi-disciplinary course titled Modeling of
Biological Systems, or MOBS. The MOBS Course is supported by the
Burroughs Wellcome Fund, and the National Science Foundation.

The goal of the MOBS Course is to bring together biologists, chemists,
engineers, mathematicians, physicists and others into a residence course to
learn from and with folks from many disciplines, and address questions of
common interests.

The MOBS 2000 Course was a rousing success for all participants, with a
positive impact on the research of the students. We are looking forward to
another exciting course in 2001.

I am pleased to call your attention to the Modeling of Biological Systems
(MOBS) Course for 2001. You will find the MOBS Course website at:
http://courses.mbl.edu/MOBS/

At the MOBS site, you will find a more detailed description of the MOBS
Course, application forms, and -- most importantly -- comments from the
students of MOBS 2000.

I encourage you to apply for the MOBS 2001 Course, and to bring MOBS to the
attention of others that may be interested in this exciting area. The MOBS
through post-docs -- and above levels of training.

I have attached pdf copies of the application and evaluation forms for

Directions for applying are on the forms and at the MOBS website. The
deadline for receipt of aplications for MOBS 2001 is December 29, 2000.

Some scholarship support is available for the successful applicants.

Should you have questions about applying to the MOBS 2001 Course -- please

Should you have technical questions about the MOBS 2001 Course, please feel
free to send them to me at: rsilver@mbl.edu

We look forward to receiving and reviewing your application.

Best wishes for a safe and enjoyable holiday season.

------------------------------

From: E.J.W. ter Maten <Jan.ter.Maten@philips.com>
Date: Wed, 06 Dec 2000 13:51:58 +0100
Subject: Conference in The Netherlands on Computing in Electrical Engineering

4th International Conference SCEE-2002
Scientific Computing in Electrical Engineering
June 23-28, 2002
Eindhoven, The Netherlands

The conference will address topics from
- Computational Elektrodynamics,
- Computational Electromagnetics and
- Circuit Simulation.

The conference will be organized by the Scientific Computing Group
of the Eindhoven University of Technology together with Philips
Research Laboratories in Eindhoven.

Further details will be announced in the future and will be available
at the Internet on http://www.win.tue.nl/scee2002/.
For subscription on information, send email to scee2002@win.tue.nl .

------------------------------

From: Chris Paul <chris@maths.man.ac.uk>
Date: Wed, 6 Dec 2000 13:26:54 GMT
Subject: Meeting in Manchester, UK on Causal Problems

The first in a series of one day meetings to be held throughout 2001 on the
analysis, applications and numerical solution of causal problems will be held
at the
Mathematics Department, Manchester University
(http://www.maths.man.ac.uk)

on

7th February 2001

The meeting will start at 10am in the 6th floor Brian Hartley conference room
with 30 minutes talks by local speakers:

Prof. C.T.H. Baker (Manchester): Linearised stability
Prof. N.J. Ford (Chester) : New numerical methods for fractional
differential equations
Dr. E. Buckwar (Manchester) : Stochastic delay differential equations
Dr. C.A.H. Paul (Manchester) : Numerical solution of delay algebraic
differential equations with non-constant lags

The afternoon session will start at 2pm with 60 minutes talks by:

Prof. D. Roose (KU Leuven) : Numerical stability and bifurcation methods in
the analysis of delay differential equations
Prof G.A. Bocharov (RAS Moscow): Numerical modelling with delay differential
equations

Any enquiries should be emailed to rfde@maths.man.ac.uk

Further details and last minute information will be made available at
http://www.maths.man.ac.uk/RFDE as it becomes available.

------------------------------

From: Yuan Ya-xiang <yyx@lsec.cc.ac.cn>
Date: Wed, 6 Dec 2000 21:29:15 +0000
Subject: Conference in China on Numerical Optimization and Linear Algebra

International Conference on Numerical Optimization and Numerical Linear Algebra
(Dunhuang, China, May 17-20, 2001)

An International Conference on Numerical Optimization and Numerical Algebra
will be held at Dunhuang, China, May 17-20, 2001. It is organized by the
Institute of Computational Mathematics and Scientific/Engineering Computing,
the State Key Laboratory of Scientific and Engineering Computing. Its sponsors
include Chinese Natural Science Foundation, Chinese Mathematical Society,
Chinese Society of Industrial and Applied Mathematics. Invited lectures on
recent advances of numerical optimization and numerical algebra will be given.

Limited number of short (about 20 minutes) contributed talks will be accepted
for presentation. Papers on theoretical, computational and practical aspects
of numerical optimization and numerical algebra are welcome. A post conference
tour to Mo-Gao Cave, a world famous visiting place, will be organized.

Prospective participants (except invited speakers) should send their
preregistration giving address (postal and e-mail, if available) and
accommodation preference(single or double room in hotel) to the address below
by post or e-mail before December 20 2000. Those who want to contribute 20
minute talks please sent the title and abstract (Latex file preferred) by
e-mail. Second announcement will be sent in January 10 2001 to those who
preregister.

Prof. Ya-xiang Yuan and Dr. Yuhong Dai
Institute of Computational Mathematics and Scientific Computing
Chinese Academy of Sciences, P.O. Box 2719, Beijing 100080, China
Tel: +86-10-255-9001, +86-10-6254-5820 FAX: +86-10-254-2485
e-mail: yyx@lsec.cc.ac.cn, dyh@lsec.cc.ac.cn

------------------------------

From: Zdenek Strakos <strakos@cs.cas.cz>
Date: Thu, 7 Dec 2000 15:24:39 +0100 (CET)
Subject: Meeting in Czech Republic on Iterative Methods

In 1997 some of us, including Ivo Marek, Miro Rozloznik, Zdenek Strakos
and Mirek Tuma on the Czech side, and Daniel Szyld, Roland Freund
and Anne Greenbaum on the US side, organized a very successful meeting
on Iterative Methods and Parallel Computing in Milovy, Czech Republic.
Based on the success of that meeting we decided to organize another one.
We are pleased to announce our second Milovy meeting called

Computational linear algebra with applications

for the period of August 4 - 10, 2002.

As to the subject, we would like to concentrate on computational
methods, namely on the methods of numerical linear algebra and their
application, e.g., to mathematical modelling, control, or image
processing.

We can accommodate about 150 - 180 participants (including accompanying
persons). We will keep the budget as low as possible; based on a
preliminary estimate the payment including the conference fee, accommodation
(double room, full board), transportation from and to Prague (Sunday
afternoon and Saturday morning), and the social program, should not
exceed \$400. Single room is possible for some modest surcharge.

The program committee consists of:

Mario Arioli (RAL, UK), Michele Benzi (Emory U., USA), Iain Duff (RAL,
UK), Michael Eiermann (TU Freiberg, Germany), Roland Freund (Lucent
Technologies, USA), Anne Greenbaum (U. of Washington, USA),
Martin Gutknecht (ETH, Switzerland), Ivo Marek (Charles U. and Czech TU,
Czech Republic), Volker Mehrmann (TU Berlin, Germany), Gerard Meurant
(CEA, France), Jim Nagy (Emory U., USA), Zdenek Strakos (ICS AS,
Czech Republic), Daniel Szyld (Temple U., USA), Henk van der Vorst
(Utrecht U., The Netherlands), Olof Widlund (Courant Institute, USA).

Mirek Tuma and Miro Rozloznik (both ICS AS) will be in charge of the local
organizing committee.

If you are interested in attending, please send a short message to
milovy@cs.cas.cz.

------------------------------

From: Maya Neytcheva <neytchev@sci.kun.nl>
Date: Thu, 07 Dec 2000 15:37:32 +0100
Subject: Conference in The Netherlands on Iterative Methods

1st announcement
Call for submissions of contributed papers to the

CONFERENCE ON
Preconditioned Robust Iterative Solution Methods
for Problems with Singularities (PRISM'2001)
University of Nijmegen, May 20-23, 2001

Special issue of NLA
An issue of the journal Numerical Linear Algebra with Applications
will be devoted to papers presented at the conference.

The topics treated during the conference include:
- construction of singularities resolving basis functions and
regularization using dominating singularity via special 3D singular elements.
- mesh refinements adapted methods: patched meshes, and successive mesh
refinement methods;
- mesh free approaches;
- defect-correction methods;
- computation of the angles between (basis) function subspaces (and
the constant in the strengthened CBS inequality) for various FEM;
- coarse mesh balanced block matrix preconditioners;
- preconditioning of the matrix block corresponding to added
basis functions (which may cause a singular behaviour);
- multigrid and algebraic multilevel iteration methods;
- iteration methods in function spaces;
- robust iterative methods for general diffusion problems and
systems of elliptic pde's including problems in elasticity (thin structures);
- inner-outer iteration methods.
- stopping criteria in iterative methods for linear and nonlinear problems.

Speakers:

N.S.Bakhvalov Moscow (RU) Jean-Francois Maitre Lyon (F)
Dietrich Braess Bochum (D) Svetozar Margenov Sofia (BG)
Evgeny Glushkov Krasnodar (RU) Ulrich Langer Linz (A)
Natalya Glushkova Krasnodar (RU) Raytcho Lazarov, College Station, TX
Piet Hemker CWI, A-dam (NL) Yvan Notay Brussels (B)
Igor Kaporin Moscow (RU) Yousef Saad, Minneapolis, Minn.
Janos Karatson Budapest (HU) Panayot Vassilevski LLNL, California
Georgij Kobelkov Moscow (RU) Freddy Wubs Groningen(NL)
Yuri Kuznetsov Houston (TX) Alex Yeremin Moscow (RU)
Ivo Marek Prague (Cz)

Scientific Committee:

Owe Axelsson (NL), Evgeny Glushkov (RU), Wolfgang Hackbusch (D),
Ulrich Langer (A), Jean-Francois Maitre (F), Svetozar Margenov (BG),
Panayot Vassilevski (USA).

Important deadline: for submission of extended abstracts (2-4 pages)
January 15, 2001.

PRISM'2001, Department of Mathematics, University of Nijmegen
Toernooiveld 1, 6525 ED Nijmegen, NL
e-mail: prism01@sci.kun.nl
URL:http://www.sci.kun.nl/math/prism01

------------------------------

From: Svetozar Margenov <margenov@cantor.bas.bg>
Date: Fri, 8 Dec 2000 15:03:02 +0200
Subject: Seminar in Bulgaria on Algorithms for Scientific Computations

SEMINAR ON ALGORITHMS FOR SCIENTIFIC COMPUTATIONS

Organizer: The meeting is organized by the Central Laboratory for Parallel
Processing (CLPP) in cooperation with the Institute of Mathematics and
Informatics(IMI), Bulgarian Academy of Sciences (BAS).

Time and place: December 18-19, room 218, CLPP-BAS, Sofia.

The event is free of registration fee, and is opened for all interested
scientists/students working in the announced area.

SCIENTIFIC PROGRAM

Monday, December 18

K. Georgiev, CLPP, Sofia, Recursive algorithms as a tool
for speed up the numerical linear algebra algorithms
Tz. Ostromsky, CLPP, Sofia, Running a large scale
air-pollution model on the EPCC supercomputers
A. Karaivanova, CLPP, Sofia, Monte Carlo and quasi-Monte
Carlo estimations for random walks on spheres and balls
E. Atanasov, CLPP, Sofia, Optimized Monte Carlo algorithm
for elastic electron backscattering from surfaces
T. Boyadjiev, Sofia University, A free boundary problem in
physics of black holes with massive dilation
M. Todorov, TU Sofia, Mathematical modeling of stars in
dilatonic gravity
L. Vulkov, University of Rousse, On the rate of convergence
of difference schemes for evolution problems with interface
J. Kandilarov, University of Rousse, The immersed interface method
for a nonlinear chemical diffusion equation with local sites of reactions

Tuesday, December 19

S. Markov, IMI, Sofia, Algebraic theory of errors
G. Dimitrov, IMI, Sofia, Three-dimensional high-order advection
terms discretization schemes. Application to the Rhone site
D. Vassileva, IMI, Sofia, Adaptive-smoothing multigrid
methods for the Navier-Stokes equations
T. Mitkova, University of Magdeburg, Germany, Concerning
numerical simulation of a magnetic-fluid seal
A. Andreev, TU Gabrovo, Superconvergence postprocessing for eigenvalues
O. Axelsson, KUN, Nijmegen, The Netherlands, On robust
preconditioning methods
M. Neytcheva, KUN, Nijmegen, The Netherlands, Algebraic solution
methods for a set of benchmark problems in linear elasticity
G. Bencheva CLPP, Sofia, MPI parallel implementation of
fast elliptic solvers
S. Margenov, CLPP, Sofia, On two preconditioning approaches
for higher order FEM systems

------------------------------

From: Rob Bisseling <Rob.Bisseling@math.uu.nl>
Date: Fri, 8 Dec 2000 16:03:03 +0100 (MET)
Subject: Euro-Par 2001 Numerical Algorithms

Call for papers
Euro-Par 2001
Manchester, UK
August 28-31, 2001

Topic 11: NUMERICAL ALGORITHMS

Topic Committee :

Global chair : Henk van der Vorst
Vice-chair : Iain Duff
Vice-chair : Bernard Philippe
Local chair : Rob Bisseling

Description :
Fast and robust parallel algorithms for the basic problems
of numerical mathematics are crucial for solving the problems in
computational science and engineering that arise today.
This workshop will be a forum for the presentation and discussion of new
developments in the field of parallel numerical algorithms,
covering all aspects from algorithmic idea, software design
and prototyping to efficient implementation on modern parallel
architectures and performance analysis. Due to its importance for HPSC
applications, the parallel iterative solution of large systems
of linear and nonlinear equations will be in the centre of interest.
Yet, contributions dealing with other problems from numerical linear
and nonlinear algebra, or dealing with general topics concerning
parallel numerical methods, are welcome, too.

Topics of interest include :
- numerical linear algebra
- large sparse or dense linear systems and eigensystems
- nonlinear systems
- fast transforms (wavelets, FFT)
- discretized partial differential equations

Euro-Par conference series :
Euro-Par is the annual European conference on parallel computing. It
is dedicated to the promotion and advancement of all aspects of parallel
computing.

For any questions related to Euro-Par 2001 please refer to our
web site:
http://europar.man.ac.uk/
or e-mail to:
europar@man.ac.uk

The key dates are:

- January 29th 2001 : Final Date for Submission
- May 1st 2001 : Acceptance Notified
- June 29th 2001 : Early Registration Deadline
- August 1st 2001 : Late Registration Deadline
- August 28-31 2001 : Conference

------------------------------

From: Philippe Guillaume <Philippe.Guillaume@gmm.insa-tlse.fr>
Date: Fri, 08 Dec 2000 18:11:43 +0100
Subject: Workshop in France on Large Sparse Linear Systems

First announce of a workshop on
LARGE SPARSE LINEAR SYSTEMS
INSA, Toulouse, France
on 1-2 march 2001

SCIENTIFIC PROGRAM

The workshop is intended to stimulate the interplay of researchers
and industrials in the field of direct or iterative methods for
the solution of large sparse linear systems, and to bring out the
actually most efficient methods as well as the most promising
research directions.

The invited lectures are:

Yves ACHDOU INSA Rennes
Abderrahmane BENDALI INSA Toulouse
Ian DUFF CERFACS Toulouse
Roland MASSON IFP Paris
Alain RIGAL Universite Paul Sabatier, Toulouse
Francois-Xavier ROUX ONERA, Paris
Yousef SAAD University of Minnesota, Minneapolis

Beside these one-hour plenary lectures, the program will include
half-hour lectures. All who are interested to contribute should
submitt an abstract.

All speakers are invited to submit an abstract in LaTeX format (to
be loaded on http://www.gmm.insa-tlse.fr/GSL2001), not exceeding two
pages, via e-mail to GSL2001@gmm.insa-tlse.fr before January 31, 2001.
Notification of acceptance of contributed talks will be sent by
february 9.

OTHER INFORMATIONS

registration, accomodation, ..) please go to the following URL:
http://www.gmm.insa-tlse.fr/GSL2001

SCIENTIFIC COMMITEE

Pierre CHARRIER Universite Bordeaux I
Philippe GUILLAUME INSA Toulouse
Mohamed MASMOUDI Universite Paul Sabatier, Toulouse
Olivier PIRONNEAU Universite Paris VI
Yousef SAAD University of Minnesota, Minneapolis

ORGANISATION COMMITEE

Luca AMODEI Universite Paul Sabatier, Toulouse
Alain HUARD INSA Toulouse
Sophie JAN Universite Paul Sabatier, Toulouse
Caroline LE CALVEZ INSA Toulouse
Benoit VERGE INSA Toulouse

------------------------------

Date: Wed, 6 Dec 2000 10:17:49 -0800
Subject: Applied Mathematics at UC Davis

APPLIED MATHEMATICS AT UC DAVIS

The Graduate Group in Applied Mathematics (GGAM) at the University
of California, Davis, invites applications from undergraduate students
interested in pursuing graduate studies (M.S and Ph.D degrees) in a strong
and vibrant program. We offer teaching and research assistantships and
fellowships to outstanding candidates. The group consists of approximately
sixty faculty members that carry out research in many areas of the
mathematical, physical, biological, chemical and engineering sciences.
The research interests of the members of the GGAM and their students
include mathematical biology, computational neuroscience, atmospheric
sciences, continuum mechanics, computational fluid dynamics, hydrology,
optimization and control, theoretical chemistry, computer and engineering
sciences, mathematical physics, signal and image processing, statistics,
combinatorial and geometric algorithms. In addition to having faculty
members who are internationally- renowned experts in applied and
computational mathematics, the GGAM includes faculty who are leading
experts in the sciences and engineering. Many of our students hold
internships at private and government research laboratories. In addition
to an exciting research environment, the University of California and the
town of Davis provide a comfortable small town atmosphere within an easy
drive of the rich recreational and cultural facilities of San Francisco,
Sacramento, and the Lake Tahoe region.

The Graduate Group in Applied Mathematics at UC Davis accepts online

http://www.math.ucdavis.edu/ggam/ or write to

Department of Mathematics
One Shields Avenue
University of California, Davis
Davis, CA 95616-8633

Please note: The GGAM requires that all applicants take the GRE to
be considered for admission. The GRE's should be taken in October or
December. The TOEFL (Test of English as a Foreign Language) score is
required for all foreign applicants whose native language is not English
or who have not studied at institutions where English is the language
of instruction. The deadline for fellowship applications is January 15th.

------------------------------

From: R. Melnik <rmelnik@mci.sdu.dk>
Date: Fri, 08 Dec 2000 02:58:34 +0100
Subject: Faculty Positions at the Mads Clausen Institute

Applications are invited for two open positions in Mathematical
Modelling and Control at The Mads Clausen Institute, University
of Southern Denmark. Further details on how to apply can be found
at our web site http://www.sdu.dk/nat/mci/vacancies/mm.html

Deadline for Applications: January 5, 2001
http://www.sdu.dk/indexE.html

------------------------------

From: Iraj Kalantari <Iraj_Kalantari@ccmail.wiu.edu>
Date: Mon, 04 Dec 2000 09:48:41 -0600
Subject: Faculty Position at Western Illinois University

TENURE-TRACK POSITION (or positions, pending approval), Assistant Professor,
August 2001. Applicants from all areas of mathematical sciences are invited
to apply. Three-course teaching (with appropriate integration of computing
technology), research, and service expected.

QUALIFICATIONS: Ph.D. (or imminent) in a mathematical sciences area;
demonstrated, or potential for, excellence in teaching; a record of, or
potential for, research; a record of, or commitment to, service.

SCREENING BEGINS December 15, 2000; continues until position filled.
Preliminary interviews at New Orleans Joint Meeting.

Send letter, vita, teaching philosophy, research description, three reference
letters, and transcripts (photocopies) to:

Iraj Kalantari, Chair
Mathematics Department
Western Illinois University
Macomb, IL 61455-1390

URL: http://www.wiu.edu/mathematics/

WIU is an Equal Opportunity and Affirmative Action employer. We are especially
interested in applications from women and minorities, and individuals with
disabilities.

------------------------------

From: Martin Gander <mgander@math.mcgill.ca>
Date: Mon, 04 Dec 2000 17:10:16 -0500
Subject: Faculty Position at McGill University

McGILL UNIVERSITY
DEPARTMENT OF MATHEMATICS & STATISTICS
Tenure Track Position in Applied Mathematics

The Department of Mathematics & Statistic at McGill university is
hiring to fill a tenure track position in applied mathematics in the
coming academic year. Preferred areas include stochastic differential
equations, material science and continuous or discrete optimization.
Candidates will be expected to have considerable expertise in
scientific computation.

Applicants should have a strong background in mathematics and have
demonstrated the capacity for independent research of excellent
quality. Selection criteria include research accomplishments, as well
as potential contributions to the educational programs of the

Applications with a curriculum vitae, a list of publications, a
research outline, an account of teaching experience, a statement on
teaching and the names, phone numbers and e-mail addresses of at least
four references (with one addressing th teaching record) should be
sent to

Professor K. GowriSankaran, Chair
Department of Mathematics and Statistics
McGill University
805 Sherbrooke Street West
Montreal, Quebec H3A 2K6
e-mail: chair@math.mcgill.ca

Candidates must arrange to have the letters of recommendation sent
directly to the above address. Candidates are encouraged to include
copies of up to 3 selected publications or preprints accepted for
publication with their applications.

McGill University is committed to equity in employment and in
accordance with Canadian immigration requirements, priority will be
applications from all candidates will be considered.

For maximum consideration applications should be received by January 15, 2001.

------------------------------

From: Homer Walker <walker@WPI.EDU>
Date: Thu, 7 Dec 2000 10:52:18 -0500 (EST)
Subject: Visiting Positions at Worcester Polytechnic Institute

VISITING ASSISTANT PROFESSORSHIPS
APPLIED OR COMPUTATIONAL MATHEMATICS
STATISTICS

The Worcester Polytechnic Institute (WPI) Department of Mathematical
Sciences invites applications for two or more anticipated visiting
assistant professorships to begin in the fall of 2001. An earned Ph.D.
or equivalent degree is required. Successful candidates must demonstrate
strong research potential and evidence of quality teaching, and will
be expected to contribute to the department's research activities and
to its innovative, project-based educational programs.

At least one position is anticipated in each of the following areas:

APPLIED OR COMPUTATIONAL MATHEMATICS. Preferred research interests
are partial differential equations with applications in fluid and
continuum mechanics, composite materials, computational modeling
and simulation, numerical analysis, parallel computing, optimization,
control theory, applied probability, and discrete mathematics.

STATISTICS. Applicants are especially encouraged in the areas of
time series, experimental design, Bayesian methods, Monte Carlo
methods, data mining methods, biostatistics, sample survey methods,
and survival analysis.

Appointments will be made for two years, contingent on satisfactory
teaching, with possible renewal for a third year at the discretion
of WPI and the appointee. The teaching load will be either five
seven-week undergraduate courses or four of these plus one
or eleven semester hours per year.

WPI is a private and highly selective technological university
and part-time graduate students. Worcester, New England's second
and recreational resources of the region.

The Mathematical Sciences Department has 24 tenured/tenure-track
faculty and supports BS, MS, and PhD programs in applied and
computational mathematics and applied statistics. For additional
information, see http://www.wpi.edu/+math.

Qualified applicants should send a detailed curriculum vitae, a brief
statement of specific teaching and research objectives, and three
letters of recommendation, at least one of which addresses teaching
potential, to either Math or Statistics Visitor Search Committee,
Mathematical Sciences Department, WPI, 100 Institute Road, Worcester,
MA 01609-2280, USA.

Applicants will be considered on a continuing basis beginning January 1,
2001 until all positions are filled.

To enrich education through diversity, WPI is an affirmative action,
equal opportunity employer.

------------------------------

From: Steve Hammond <hammond@ucar.edu>
Date: Thu, 7 Dec 2000 09:57:27 -0700
Subject: Position at the National Center for Atmospheric Research

Scientific Programming/Software Engineer at NCAR

BASIC FUNCTION OF JOB:

We invite applications for a position that will:
* Develop efficient scalable software and algorithms for use in
geoscience applications on parallel computers using message
passing and multi-processing mechanisms such as MPI and OpenMP.
* Conduct simulation of fundamental problems in turbulence on
state-of-the-art computer systems.
* Develop adaptive mesh refinement techniques.
* Collaborate with NCAR staff on efficient use of high
performance parallel computers and analysis tools.
* Primary effort within the first 2 years will be development of
a new turbulence code.

REQUIRED QUALIFICATIONS INCLUDE:

* Masters of Science in computer science, physics, fluid dynamics
and MHD, applied math or a related field or the equivalent to
completion of a M.S.
* Experience with simulations that use parallel code on
supercomputers as well as the workstation environment.
* Ability to interact with a broad scientific effort that
features large simulation projects.
* Knowledge of contemporary software engineering techniques.
* Working knowledge of numerical analysis, in particular in
dealing with PDEs, and with data analysis.
* Demonstrated skills in C, C++, and Fortran.
* Demonstrated skill in parallel algorithm development:
distributed memory using MPI and/or multithreading for shared
memory programming using OpenMP or pthreads.
* Familiarity with measuring and evaluating performance and
scaling of parallel algorithms and developing efficient codes
for RISC microprocessors.
* Knowledge of visualization techniques as applied to the
analysis of large data sets.

The successful candidate will participate in the development of new
simulation methods in collaboration with research teams at
the National Center for Atmospheric Research, will assure
that the codes are performance portable, and will analyze
results using contemporary visualization techniques.

http://www.fin.ucar.edu/hr/employment/1024.html

------------------------------

From: Thomas Harte <Thomas.Harte@kla-tencor.com>
Date: Mon, 4 Dec 2000 16:32:15 -0800
Subject: Position at KLA-Tencor, Milpitas, California

Open requisition #16197: Senior Software Engineer

The KLA-Tencor Algorithms Group in Milpitas, California, is offering a
Senior Software Engineer post to:

- optimize C/C++ to advanced algorithm design specifications
for execution on a multi-processor architecture

- develop algorithm benchmarks in low-level C/C++ for

The successful candidate will enjoy the challenge of developing
low-level, optimized advanced algorithm code in a time-critical
environment and with cutting-edge computer architectures. MS/PhD in
computer engineering, EE or related discipline. Minimum 2 years
experience of C/C++ in an image processing or advanced algorithm
environment. The candidate should have a strong interest in UNIX systems.

Applicants must be resident in the United States or must hold United States
residency visa status.

Contact: Thomas.Harte@kla-tencor.com

------------------------------

From: Dan Hitchcock <hitchcoc@er.doe.gov>
Date: Tue, 05 Dec 2000 13:09:44 -0500
Subject: Program Manager Position at Department of Energy

The Mathematical, Information, and Computational Sciences (MICS) Division,
Office of Advanced Scientific Computing Research (ASCR), Office of Science
(SC), US Department of Energy is seeking an applied mathematician to serve
as Program Manager for Applied Mathematical sciences. This program focuses
on multidisciplinary applied mathematics research activities that benefit
the Department's ongoing science and technology program. This person would
come to DOE under the provisions of an Interagency Personnel
Agreement. For this reason applicants should be currently full time
employees of educational or other non-profit organizations.

The Office of Advanced Scientific Computing Research provides a focal point
in the Office of Science for long-term computational mathematics research.
The mission of this office is to discover, develop, and deploy the
computational and networking tools that enable researchers to the
scientific disciplines to analyze, model, simulate, and predict complex
phenomena important to the Department of Energy. The Office fosters and
supports fundamental research in advanced scientific computing - applied
mathematics, computer science, and networking - and operates supercomputer,
networking, and related facilities. The applied mathematics program has
made significant investments in areas such as computational fluid dynamics,
numerical linear algebra, computational methods for ode's and pde's, and
large scale optimization, as well as other areas which have the potential
for significant impact on DOE's missions.

The program manager for applied mathematics serves as a recognized
technical authority and expert in the Department in the field of applied
mathematics relating to computational science; determines pioneering
research needs and opportunities of this research program against
scientific and technological advances and potential needs of the
Department; manages the review and execution of research proposals from
universities and national laboratories; and serves as a critical
communications link between the Department of Energy, other Federal
Agencies, and the broad research community.

Research, at (301)903-6767, or hitchcock@er.doe.gov.

------------------------------

From: T. Terlaky <terlaky@mcmaster.ca>
Date: Thu, 07 Dec 2000 23:02:15 -0500
Subject: Postdoctoral Position at McMaster University

(http://www.cas.mcmaster.ca/~oplab)
of the Department of Computing and Software,
McMaster University has one postdoctoral position available.

The successful applicant should have a Ph.D. in Operations Research,
(Computational) Mathematics or Computer Science, with
strong background in at least one of the following areas:
algorithms, optimization, mathematical programming,
scientific computing.
The research project will focus on implementing, benchmarking
and further developing the recently discovered Self-regular
proximity based Interior Point Methods. For more details see
http://www.cas.mcmaster.ca/~oplab/IPMs/newipm.html

Applicants should send a CV to
Prof. Tam=E1s Terlaky
terlaky@mcmaster.ca
possibly as a Postscript/PDF file, as soon as possible.

The successful applicant will work in close cooperation
with the New IPMs Research Group.
Starting date, with mutual agreement, is as soon as possible.
The salary is approximately \$30000.- CAD + health insurance benefits.

Further information can be obtained from:
Prof. Tam=E1s Terlaky
Tel: +1-905-525 9140 ext: 27780
terlaky@mcmaster.ca
http://www.cas.mcmaster.ca/~terlaky/

------------------------------

From: Fred Dulles <dulles@ima.umn.edu>
Date: Mon, 04 Dec 2000 14:21:40 -0600
Subject: Postdoctoral Positions at University of Minnesota

The Institute for Mathematics and its Applications

ANNOUNCES TWO YEAR POSTDOCTORAL MEMBERSHIPS

in the 2001-2001 Annual Program
MATHEMATICS IN THE GEOSCIENCES
http://www.ima.umn.edu/geoscience/

ORGANIZING COMMITTEE : William I. Newman (Chair), David R. Brillinger,
Michael Ghil, J.M. Hyman, Frederic Schoenberg, William W. Symes,
Donald L. Turcotte, Mary F. Wheeler

A one-year program with three parts:

Fall: September - December 2001, Dynamical Systems and Ergodic Theory

Winter: January - March 2002, Multiscale Phenomena and Renormalization

Spring: April - June 2002, Inverse Problems and the Quantification
of Uncertainty

Further information including requirements, applications and details
about the program may be found at http://www.ima.umn.edu/docs/forms.html
and at the first URL above or by inquiry to ima-staff@ima.umn.edu or
612/624-6066. The application deadline is January 15, 2001 for receit of
all material.

Information about Senior Memberships and Industrial Postdoctoral
positions is also available from the above sources.

------------------------------

From: Bob Russell <rdr@cs.sfu.ca>
Date: Mon, 4 Dec 2000 15:13:21 -0800 (PST)
Subject: Postdoctoral Positions at Simon Fraser University

Postdoctoral Fellowships
Applied and Computational Mathematics Group
Department of Mathematics and Statistics
Simon Fraser University

The Applied and Computational Mathematics Group of Simon Fraser University
invites applications for postdoctoral positions in numerical PDEs. We seek
promising researchers, generally within 3 years of their PhD, who have
particular expertise in some subset of spectral methods, integral equation
methods, adaptive grid techniques, level set methods, or finite
element/difference methods.

The positions may be partially funded through the Pacific Institute of
Mathematical Sciences (PIMS) and will be research positions with a maximum
of one course of teaching a year. There is some flexibility on when the
postdoctoral position would commence, although a PIMS position would
normally begin between April 1, 2001 and January 1, 2002. (For details on
PIMS, see http://wren.pims.math.ca/). The positions would generally have
two years duration.

Simon Fraser University has a young and very dynamic Applied and Computational
Mathematics Group. A postdoctoral position at SFU would offer an enjoyable
working environment and unique physical surroundings. See
http://www.math.sfu.ca/applmath/ for further information on our group and our
research interests.

There are no citizenship restrictions on the positions. Interested applicants
can contact Bob Russell by email (rdr@cs.sfu.ca) for further inquiry or can
directly send two letters of reference, a curriculum vitae, and a statement
of research interests to Bob Russell, Department of Mathematics and
Statistics, Simon Fraser University, Burnaby, BC, Canada V5A 1S6 by January
22, 2001.

------------------------------

From: Iain Boyd <iainboyd@engin.umich.edu>
Date: Wed, 6 Dec 2000 15:50:08 -0500 (EST)
Subject: Postdoctoral Position at University of Michigan

Post-Doctoral Research Associate
Particle Simulation of Plasmas
Department of Aerospace Engineering
University of Michigan, Ann Arbor, MI 48109.

An immediate opening is available for a post-doctoral research
associate in the area of developing particle methods for simulation
of plasmas. Initial work would focus on developing algorithms to
reduce statistical fluctuations in simulations of electron streams.
Additional on-going projects are concerned with particle simulations
of spacecraft plasma propulsion systems.

A candidate is sought with a strong background in: (1) particle
simulation techniques; and (2) plasma physics.

The position is available immediately, and is guaranteed initially
for one year.

Contact: Professor Iain D. Boyd
Mail: University of Michigan, 3012 FXB Building,
Ann Arbor, MI, 48109-2140.
Telephone: (734) 615-3281
FAX: (734) 763-0578
E-mail: iainboyd@umich.edu
Web: hpcc.engin.umich.edu/CFD/research/NGPD

------------------------------

From: Einar Ronquist <ronquist@math.ntnu.no>
Date: Thu, 7 Dec 2000 14:04:15 +0100
Subject: Postdoctoral Position at NTNU, Norway

The Department of Mathematical Sciences at the Norwegian University
of Science and Technology (NTNU) in Trondheim, Norway, announces a
new postdoctoral position in applied mathematics. The position is
funded by the Norwegian Research Council (NFR) over a period of 1 year.

The successful candidate should work with numerical methods for
simulating fluid flow with small density variations, but with
potentially large density gradients. The methods should be applied
to physical situations where internal waves are important.

The successful candidate will become part a research group involved
in the project "Computational methods for stratified flows involving
internal waves."

Applicants must have a doctoral degree in mathematics, informatics,
or other relevant background in computational methods.
Doctoral candidates who expect to finish their degrees
during the academic year 2000/2001 are encouraged to apply.
The deadline for application is December 31, 2000.

Questions regarding the opening can be directed to:
Einar Ronquist: ronquist@math.ntnu.no, or to
Brynjulf Owren: Brynjulf.Owren@math.ntnu.no

For further informasjon about the project and the vacant position,