Pages: 1
Author Topic: ENIGMA VS Game Maker (Speed & Graphics)  (Read 5506 times)
Offline (Unknown gender) Lemon Pie
Posted on: April 26, 2012, 08:37:30 pm
Joined: Mar 2012
Posts: 9

View Profile
I made a project in Game Maker that would draw a Mandelbrot Set fractal (depending on a few parameters). I ported the project to ENIGMA to test speeds of ENIGMA vs GM.

Visually this is what happened  :o

(ENIGMA on the left, Game Maker on the right)

Speed-wise I got a 10-30 times increase in ENIGMA over GM :D
I was able to render a fractal in ENIGMA at 50 iterations per x,y coordinate about 10 times in the time it took for GM.

Anyways..  :v:
I can't wait to see how far the ENIGMA project goes! Great work so far guys!  (Y)

The Mandelbrot Set script in GML for anyone who may want it:
Code: [Select]
//Mandelbrot set algorithm//
draw_set_color(c_black)     //sets drawing color to black
draw_rectangle(0,0,room_width,room_height,false)    //draws 'background'
screen_refresh();           //refreshes screen
xx = 0; //starting x point
yy = 0; //starting y point

Iteration = global._Iterations; //Maximum Iterations (Number of times it will redraw)
MinReal = global._MinReal;      //Minimum Real Number (Smallest, Negative)
MaxReal = global._MaxReal;      //Maximum Real Number (Largest, Positive)
MinIm = global._MinImaginary;   //Minimum Imaginary (Smallest Imaginary, Negative)
MaxIm = MinIm+(MaxReal-MinReal)*room_height/room_width;
                                //Maximum Imaginary (Largest Imaginary, Depends on Others)

RealFac = (MaxReal-MinReal)/(room_width-1); //Real factor
ImFac = (MaxIm-MinIm)/(room_height-1);      //Imaginary Factor

cReal = MinReal + xx *RealFac;  //C real start value
cIm = MinIm - yy * ImFac;       //C Imaginary start value

for (yy = 0; yy <room_height; yy+=1)
    cIm = MaxIm-yy*ImFac;   //New C Imaginary value
    for(xx = 0; xx<room_width; xx+=1)
        cReal = MinReal+xx*RealFac; //New C Real value
        zReal = cReal;  //Z Real value
        zIm = cIm;      //Z Imaginary value
        isTrue = true;
        for (i=0; i<Iteration; i+=1)
            zR2 = sqr(zReal);   //New z[Real]^2
            zIm2 = sqr(zIm);    //New z[Imaginary]^2
            if (zR2+zIm2>4)     //if the 2 numbers are inside the fractal's limits
                isTrue = false; //sets isTrue to false
                if argument0 = true //if you want to draw a background color
                    draw_set_color(make_color_hsv(zR2*(5*i),0,i*10))  //makes a color
                    draw_point(xx,yy)   //draws the color at the point
            zIm = 2*zReal*zIm+cIm;  //New z[Imaginary]^2
            zReal = zR2-zIm2+cReal; //New z[Real]^2
        if (isTrue) //if the point should draw
            if argument1 = true //if you want to draw inside the fractal image
            {   //makes a pretty color for the inside of the fractal
            else draw_set_color(c_white)    //else the color is white
            draw_point(xx,yy)   //draws a pixel at XX,YY w/ color           
    screen_refresh();   //refreshes the screen so you can see the fractal creating itself
                        //Putting the refresh here is faster than after a pixel is drawn.
    sleep(1);           //sleeps for 1 millisecond. Allows you to
                        //keep control of the window while drawing in GM
Offline (Male) polygone
Reply #1 Posted on: April 26, 2012, 09:16:35 pm

Location: England
Joined: Mar 2009
Posts: 794

View Profile
There's some more games on the EDC if you want to test the speed of GM vs ENIGMA. These two are good:
I honestly don't know wtf I'm talking about but hopefully I can muddle my way through.
Post made April 27, 2012, 02:23:00 am was deleted at the author's request.
Offline (Unknown gender) TheExDeus
Reply #3 Posted on: April 27, 2012, 02:48:13 am

Joined: Apr 2008
Posts: 1860

View Profile
The same reason ray trace demo from EDC look interlaced. I suspect it was flashing (like switching between the lines) right? I never could of figure out why this happened. I thought it was because of buffer switching during screen_refresh() but I am not sure. Anyway, as surfaces are in, you could probably use those to eliminate these problems.

edit: Just tried the code. Have the same problem. I fixed it by using surfaces and screen_redraw().
SwapBuffers(enigma::window_hDC); (on win) is the cause of the problem and I don't think you can easily fix it.

edit2: Its something to do with WM_ERASEBKGND message and OnEraseBkgnd. If they are overwritten then it could fix it (hell if I know how to overwrite them).
« Last Edit: April 27, 2012, 03:45:30 am by HaRRiKiRi » Logged
Offline (Male) Josh @ Dreamland
Reply #4 Posted on: April 27, 2012, 11:16:09 am

Prince of all Goldfish
Location: Pittsburgh, PA, USA
Joined: Feb 2008
Posts: 2950

View Profile Email
It is because of buffer switching. Half of the lines are painted on the front buffer, half on the back. Just draw it all in one go. :P

Perhaps we could use for an option to disable double buffering...

At any rate, it's nice to see someone able to get the project running without hassle. :)

Also, you could speed this up more by placing it in draw_primitive_begin(pr_pointlist)/draw_primitive_end(), and using draw_vertex() instead of draw_point().
« Last Edit: April 27, 2012, 11:20:17 am by Josh @ Dreamland » Logged
"That is the single most cryptic piece of code I have ever seen." -Master PobbleWobble
"I disapprove of what you say, but I will defend to the death your right to say it." -Evelyn Beatrice Hall, Friends of Voltaire
Offline (Unknown gender) TheExDeus
Reply #5 Posted on: April 28, 2012, 03:11:59 am

Joined: Apr 2008
Posts: 1860

View Profile
But linux also has double buffering and you don't see this interlacing. I also thought it was something to do with double buffering (like half a year ago when I clearly mentioned this problem), but googling makes me to believe its something to do with windows and WM_ERASEBKGND. Its apparently a common problem with gl on windows.

edit: Also, OP if you used data types that are in ENIGMA, then you should be able to speed it up even more. Like putting int in front of some variables.
« Last Edit: April 28, 2012, 03:14:34 am by HaRRiKiRi » Logged
Pages: 1