The Mandelbrot fractal is quite a sight.
Such wonder and amazement that can be found while viewing the Mandelbrot fractal can now be had on the Casio Prizm graphing calculator.
In addition this program can also be compiled for the PC assuming SDL is installed.
Just run make and it will be up and running.
There is very little difference in function between the PC version and the casio prizm version, the resolution is higher but that is it.

The controls are
Left, right, up, down moves the window in that direction
Menu key or ESC on the PC, exits to the menu or exits the program (PC)
F1 sets maximum iterations to 65535
F2 sets the maximum iterations to 224 (this is the default upon starting the program)
F3 toggles deep mode which is on by default. This means you can zoom in deeper with a slight performance hit.
I do not notice the difference
1 subtracts one from the maximum iterations
2 adds one to the maximum iterations
3 subtracts ten from the maximum iterations
4 adds ten to the maximum iterations
5 subtracts 100 from the maximum iterations
6 adds 100 to the maximum iterations
Shift zooms in
Alpha (ALT on PC) zooms out
Source code https://github.com/ComputerNerd/Mandelbrot-Casio-Prizm-Explorer
Binary https://github.com/ComputerNerd/Mandelbrot-Casio-Prizm-Explorer/blob/master/mandelbrot.g3a
Screen shots (from the PC version which gets very similar output differing only in resolution).


Anyway if you have any bug report(s), feature request(s), pull request(s), patch(es) or whatever relates to this I would be happy to hear about them.

I have to say these shots look really cool. I need to dump this on my Prizm and check it out

Yes just a tip, to get those nice deep blue colors I pressed F1, which increases the maximum iterations and also affects color. Also your post suggests that you don't have your prizm handy. You can also run this on your PC to get a feel for what you will get on the casio prizm.

This program has been updated to support the CG-50. The link to download is the same: https://github.com/ComputerNerd/Mandelbrot-Casio-Prizm-Explorer/blob/master/mandelbrot.g3a

Note that this is untested. Let me know if it works.

All that I get is a white screen on a CG50

This is my fault.
In my linker script I had

Code:

Which works on the CG-10/20 but not on the CG-50.
This

Code:

should work on both models.

Please re-download and let me know if this fixed the issue.

I now get a black screen with a white border of about 6 pixels.
(CG50)

same for me unfortunately on fx-cg50 hardware

I recompiled it with the latest version of GCC. Maybe it will work now.

https://github.com/ComputerNerd/Mandelbrot-Casio-Prizm-Explorer

It works on my CG10.

I tried on the CG50.
It stops at the black screen in a white frame.
It is possible to return to MENU without freezing.

Here is my latest attempt of getting it working on the fx-CG50. I could not see anything in the code that would cause that issue and given that I don't own a fx-CG50 the best I could do was recompile with the latest version of GCC and Binutils and hope for the best: https://github.com/ComputerNerd/Mandelbrot-Casio-Prizm-Explorer/blob/master/mandelbrot.g3a

This is display.h is not fx-CG 50-compatible.


Code:

On the fx-CG 50 the RAM starts at 0c000000 in the physical address space. I can see you use the syscall to get it, but both VRAM_ADDR and VRAM_ADDRESS are used interchangeably in main.c.

I tried to test it (I'm very confident it would work immediately) but I've got an ld internal error and I don't want to debug it right now. If you can make another build I'll be happy to try it

Sorry for missing that. I resolved the issue and I've updated the G3A. The link is the same. Please try again.

It works now! This program is nuts. I once made a prototype that could draw the whole fractal decently fast (0.3 ms?) but I never thought of doing progressive detailing, which helps a lot keeping the thing fluid at deeper levels.

Thank you for testing. I'm glad it's now working on the fx-CG50.

Also what unit is that for the time? If it's milliseconds that would be over 3333 fps (1000 / .3). Would it perhaps be 0.3 seconds (1 / .3 = 3.3... fps)?

Yes it was 0.3 seconds, 3000 FPS is straight up impossible lol. Good catch. I was using 32-bit fixed point for the outer region (as low precision is sufficient) with hand-written assembler and the whole 396x224 image rendered in 311 ms. I thought it was not bad, even though 32-bit reaches its limits rather quickly.

I would probably have expanded it into a whole viewer on a rainy Sunday eventually, but now that I've seen your program it wouldn't be as interesting. ^^