|
meta::Expr<T_emplate>
Easily add expressions templates to your classes in C++ (1998-2000)
meta::Expr<T_emplate>release 15 Dec 99
(meta-1.17.tar.gz)
[status: 1.17 release]
This module provides an easy way to add Expressions Templates to your C++
classes. You will gain in a lot performance. This is due to a better
optimization: the compiler to do not generate temporary variable in
simple cases anymore.
Download
(meta-1.17.tar.gz)
Look at the source
Documentation
autodiff::Fad<FLOAT>
Automatic differentiation in forward mode (1995-1999)
autodiff::Fad<FLOAT>
(autodiff-3.04.ps)
(autodiff-3.04.pdf)
(autodiff-3.04.tar.gz)
[status: 3.04 beta]
This modules provides automatic differentiation classes in forward
mode. Performance are much better than in
ADOL-C
or in
FADBAD.
SL++: a C++ library for doing numerical analysis
provides complex, quaternion, matrix and vector classes(1997-1999)
-
Official nice Home Page
- SL++ The library is under the GPL.
- Installation
(html)
(ps)
[status: alpha]
Please read this pages in order to install the library.
- Module Matrix
(html)
(ps)
[status: 2.10]
This modules provides vector and matrix manipulations. It is
currently near a BLAS3 + SPARSEBLAS library.
- Module Complex
(html)
(ps)
[status: 1.1]
This is a rewrite of std::complex with some functions added.
- Module Quaternion
(html)
(ps)
[status: 1.1]
This is an implementation of quaternions. Conversion with 3D
rotations matrices is provided.
- Module Algebra
(html)
[status: alpha]
This is module is not usefull at this state.
- Module Misc
(html)
[status: alpha]
This modules provides a lot of simple usefull functions for
benchmarking or dealing with compiler oddities.
- Scientific Library Mailing Lists Archive
- Scientific Library Download (last stable version)
- Scientific Library Download (last unstable version)
- Associated Publications
- P. Aubert, N. DiCesare, O. Pironneau. Automatic Differentiation in C++
using expressions templates. Submitted to Computer Software Vision.
C-optim
A C-finite element program for shape optimization (1993-1997)
- Features:
- Geometrically nonlinear beam (2d and 3d elements, model from Fried, Simo and HibraimBegovic)
- Geometrically linear plates (2d)
- Geometrically (non)-linear shells (3d)
- Thickness optimisation
- Shape optimisation
- Code is available for free for university for research or teaching purpose.
Coptim use a lot of other softwares: LAPACK, ARPACK, PETSc, CFSQP, M1Q3N,
GNUPLOT, PLPLOT, MAPLE V, SED, AWK ...
- Associated publications:
- Pierre
Aubert and Eric Hassold.
Comparison between formal calculus and automatic differentiation in the context
of shape optimization.
Technical report, Université de Nice-Sophia Antipolis, Nice, France,
1995.
- Pierre
Aubert and Bernard Rousselet.
Sensitivity analysis of a geometrically nonlinear 3d arch.
submitted to journal of applied and theoretical optimization,
1996.
- Pierre
Aubert and Bernard Rousselet.
Sensitivity computations and shape optimization for a nonlinear arch model with
simple instabilities.
International Journal for numerical methods in engineering,
42:15-48, 1998.
- Pierre Aubert.
Comparison between automatic differentiation and formal calculus in optimal
design application.
In Control and Cibernetics, editors, Shape optimal design, theory,
applications and software, volume 1. CIMPA/G.D.R./TEMPUS-JEP intensive
school, 1995.
- Pierre Aubert.
Une méthode d'optimisation de forme en présence d'instabilités.
In Deuxième colloque national en calcul des structures,
volume 2, pages 795-800. Association Calcul des Structures et
Modélisation, Hermes, May 1995.
- Pierre Aubert.
Optimisation de forme en présence d'instabilités.
PhD thesis, Université de Nice-Sophia Antipolis, Nice, France, Laboratoire
J.A. Dieudonné, Parc Valrose, 06108 Nice Cedex 2, 1996.
- Pierre Aubert.
Sensitivity analysis for a geometrically nonlinear arch model.
In J.-A. Desideri, P. Le Taller, E. Oate, J. Periaux, and E. Stein, editors,
Numerical methods in Ingeneering, pages 256-263. Eccomas,
Wiley, september 1996.
- Pierre Aubert.
Thickness optimization of a geometrically non linear arch at limit points.
Applied Mathematics and Computer Science, 6(2):101-117, 1996.
Interp
A small interpreter (1992-1993)
- Features: parse a string and return a way to evaluate this
expression
For example:
string="X*sin(Y)+1"
can easily be evaluated in X and Y
- Source use lex and yacc and is available under the GPL.
Look for the source here or a .tar.gz is
available here.
- I wrote Interp in 1993. I don't claim that is work but it is fairly
ANSI-C code.
|