dpalahn@guido.uucp (Donald Palahnuk) (07/06/89)
Sorry to have to repost this in this news group
but E-Mail has been bouncing as usual,
ATTENTION JERRY LEE :: HERE IS THE CODE YOU REQUESTED.
Included is addended code that allows variable view size !
notes:
this computation produces the Julia Set of Lambda*ComplexCosine(Z)
the algorithm was developed from RL Devaney's Hubbard Alg apprch.
super major godlike disclaimer:::: works on a SUN 4/110, don't
ask me what else, period!!!
Although I would love any theory questions on the alg,
I'm a mathematician and chemical eng'r not a comp sci'tist.
However, what I am aware of is that this code may only
be SUN4 compatible???
see further notes below, donnie
IMPORTANT NOTE: you must extract the following code from this file.
Following are three sets of code which all must be extracted.
1. go_var: gofile to comp, link and exe.
2. c_var.c: graphics source code
3. c_var.F: computational source code. (note CAP F extension
required for CC precompilation! )
notes: choose 75 for the starting array size to see if it works for speed
lambda is the funky parameter: change it to change the image
1. go_var
echo comp f77 c_var
f77 -c c_var.F -o cf_var.o
echo comp cc c_var
cc c_var.c cf_var.o -o cfor_var -lpixrect -lsunwindow -lsuntool -lm -lF77 -lI77 -lU77
echo execute cfor_var
cfor_var
notes: The library extensions -lF77 -lI77 -lU77 are optional (I think)
and are only required if you require i/o from the fortran subroutine
level and don't want the infamous core dump.
The file.F for f77 should be used for speed and optimization.
2. c_var.c
# include <suntool/sunview.h>
# include <suntool/canvas.h>
# include <stdio.h>
# include <math.h>
#include <sunwindow/cms_rgb.h>
#define CMS_SIZE 64
/* graphics routine for donnie's fractal image prog */
u_char red[64] = { 0,100,130,200,225,255,255,255,255,255,255,
/* 11*/ 255,255,255,255,255,255,255,240,223,193,168,140,120,100, 45, 10,
/* 27*/ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
/* 43*/ 0, 0, 0, 0, 0, 0, 0, 0, 0, 75, 96,114,135,150,168,193,
/* 59*/ 218,255,255,255,255};
u_char blue[64] = { 0, 0, 0, 0, 0, 0, 60, 90,100,120,138,
/* 11*/ 151,168,193,215,230,245,255,255,255,255,255,255,255,255,255,255,
/* 27*/ 255,255,255,255,255,255,255,255,255,255,255,255,255,243,229,205,
/* 43*/ 189,175,161,139,122,105, 69, 35, 0, 0, 0, 0, 0, 0, 0, 0,
/* 59*/ 0, 0,173,233,255};
u_char green[64] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
/* 11*/ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
/* 27*/ 0, 50, 95,120,145,160,175,191,205,218,231,241,255,255,255,255,
/* 43*/ 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
/* 59*/ 255,255,255,255,255};
main()
{ /* graphics attributes variabke declerations */
Frame frame;
Canvas canvas;
register Pixwin *pw;
int i,j,quib;
int flag[500][500];
int tag;
printf("enter array dimension\n");
scanf("%d",&quib);
printf("%d\n",quib);
/* create canvas and frame */
frame=window_create(NULL,FRAME,0);
canvas=window_create(frame,CANVAS,0);
pw=canvas_pixwin(canvas);
/* initilize a roy-gee-biv */
/* send info to pixwindow */
pw_setcmsname(pw,"showcolor");
pw_putcolormap(pw,0,CMS_SIZE,red,blue,green);
/* get information for plotting */
/* note that the transfer below are pointed addresses ! */
res_(flag,&quib);
/* send information to window */
for(i=0;i<=quib-2;++i){
for(j=0;j<=quib-2;++j){
pw_rop(pw,i+10,j+10,1,1,PIX_SRC | PIX_COLOR(flag[i][j]*2),
(Pixrect *)0,0,0);
}}
window_main_loop(frame);
exit(0);
}
notes: VERY IMPORTANT (or fairly anyway), transfer of variables
to fortran code must be pointed addresses ! Look @ res_(flag,&quib) !
Secondly the program has an upper limit of 500X500 pixels, this may
need to be reduced if you run into memory problems. To make an image
larger the computation would need to be done in segments.
3. c_var.F
c Complex Plane Iterative Routine
c As Per R. Devaney and the Hubbard Algorithm
function res(flag,quib)
c give it a try
integer i,j,k,flag(500,500),quib
real bound
complex z
complex*16 zt
real im,Pi
real lambda
Pi=3.141592654
bound=quib/2.0
lambda=2.94
do i=1,quib
do j=1,quib
k=0
z=cmplx((bound-i)*Pi/(2*quib),(bound-j)*Pi/(2*quib))
do while (flag(i,j).eq.0.and.k.lt.31)
k=k+1
zt=lambda*ccos(z)
im=dimag(zt)
if(abs(im).gt.25) then
flag(i,j)=31-k
endif
z=zt
enddo
enddo
enddo
return
end
notes: I don't think I have any for this section, besides the I/O mentioned
above concernig the proper libraries. (which may or may not be suppported
by SUNS below model 4?)
Thank you for your interest, donnie