Three programs were done to study the attractor, all of them in ForTran,
writing results in matlab files.
To Download the code click here
The Three Programs : User's Guide
The first program is obtained by compiling calcul_unique.f. It does the
computation of the serie for one set of parameters asked just after the
program is launched.
The second program, obtained after compilation of calcul_tous.f does
the same computation, but for a range of lambda and beta specified.
The third program, called master and slave, are used after using make.
They need PVM to run. It does the same thing as the second program, but
using n processors, and writing n files at a time.
All programs can export Matlab files or standard text files in the format
"x y". Matlab files are executed ubder Matlab abd automatically plot the
result, storing the coordinates of the points in a matrix A, and a matrix
B.
Known bugs :
*The program tends to crash if asked to calculate very large simulations
(over 30000 for calcul_unique, and over 10000 for the parallel version)
The codes
Direct computing
The algorithm is very simple. It is a loop over the number of iterations
asked by the user, storing the coordinates of the series in two vectors.
The vectors, once build are then stored in a result file.
Compute many parameters
The algorithm is a modified version of the direct computing program. It
loops over the direct computing program, feeding it with different lambda
and betas, in a standard loop :
do while beta less than beta max
lambda=lambda start
do while lambda less than lambda max
calculate the serie
increase lambda
end do
increase beta
end do
Parallel computation on a set of parameters
Programmed under PVM, this program uses the initialization given in a program
developed by Mr. Dayde. The rest was written specially for this task. It
consists in distrtibuting the direct computing of one parameter to many
processes. This allows a gain of time, compared to the second program,
but is limited by the speed of the Hard disk of the computer used for storage,
since many files are saved at the same time.