fluid64 by Rudi
screenshot added by rudi on 2014-09-22 20:04:57
platform :
type :
release date : september 2014
  • rulez 11
  • is ok 3
  • sucks 0
popularity : 54%
  • rulez 0.79
alltime top: #10971
added on the 2014-09-22 20:04:57 by rudi rudi

popularity helper

increase the popularity of this prod by spreading this URL:

or via: facebook twitter pinterest tumblr


yesterday i found some ol' code from march 2013 and decided to crush the bytes down to the 64-byte limit.
let it run for a while, and the particles will form spiral-vortices and waves. the state of the system tries to balance itself but it will never reach equilibrium.
hope you like it.
added on the 2014-09-22 20:06:08 by rudi rudi
neat :)
rulez added on the 2014-09-22 20:07:50 by sensenstahl sensenstahl
rulez added on the 2014-09-22 20:27:28 by whizart whizart
rulez added on the 2014-09-22 20:32:16 by Starchaser Starchaser
Locks up with gfx crap here, what is it tested working on?
added on the 2014-09-22 20:48:07 by Photon Photon
Photon: I'ts primarily coded and tested on Dosbox 0.74 machine=vgaonly, speed=approx 200k cycles (~full). But i've tested it on a Compaq with dual core 1.6GhZ hardware and DOS 6.22 installed.
added on the 2014-09-22 21:08:40 by rudi rudi
rulez added on the 2014-09-22 21:22:11 by BiTL BiTL
reminds me HodgePodge;) great
rulez added on the 2014-09-23 06:00:33 by g0blinish g0blinish
Looks a bit different here. Different Memory/register defaults, maybe?

BB Image
added on the 2014-09-23 11:11:41 by tomaes tomaes
rulez added on the 2014-09-23 11:53:30 by Optimonk Optimonk
Sweet =) I crunched it down to 61 byte, and changed the colors a bit

BB Image

Also did an escapeable version in 64 byte, but that's very slow and looks different.


I think, this can be brought down way further in size, but well, this already exists ;)
rulez added on the 2014-09-23 16:21:02 by HellMood HellMood
Forgot to say, all versions - including the original - may behave different on different machines. There is a blind write to a whole segment adressed by BP which can point to OS data and whatnot (see here ). Also, when it's completely empty, nothing happens. So, from black (blue) screen to system crash, everything is possible ;)
added on the 2014-09-23 16:25:42 by HellMood HellMood
b013 cd10 c40f 8edd b100 bd08 008b 9e35
0102 09f7 db02 094d 4d75 f2c0 e902 8a05
4824 1f28 c83c 0a7d 0a24 01f6 d802 0524
1f88 05ac aaeb d13f 0140 0141 0101 00

And this is the entire code in hex ^^
rulez added on the 2014-09-23 17:18:30 by The_Grandmother The_Grandmother
tomaes: yep, as Hellmood pointed out, the segment address can point to different places on different systems.

Hellmood: Řrřola use the Von Neumann neighbourhood in Dírojed. I use the Moore neighbourhood. This alone increases the size. I had to fiddle about the arithmetic and fine-tune to produce those spirals. Yes, I bet there is some room to strip even more bytes off. But I am still not very good at that.

and basically the Turbo-C version I wrote first (if I found the right code, on paper) looks like this:

Code:for (i=640; i<63360; i++) { c=buf[i-321]; c+=buf[i-320]; c+=buf[i-319]; c+=buf[i-1]; c+=buf[i+1]; c+=buf[i+319]; c+=buf[i+320]; c+=buf[i+321]; 128-(c>>3)+((buf[i]-1)&0x3f)<169?buf[i]=(buf[i]-(r&1))&0x3f :0 r+=7; asm rol,1 b+=c; r^=b; }

where buf is the noise buffer (initial configuration).
and more close to this version:

Code:for (i=640; i<63360; i++) { 128-((v[i-321]+v[i-320]+v[i-319]+v[i-1]+v[i+1]+v[i+319]+v[i+320]+v[i+321])>&g t;1)+ ((v[i]-1)&15)<122?v[i]=15&v[i]-(r&1):0; r+=n[i]; }
added on the 2014-09-23 23:12:03 by rudi rudi
rulez added on the 2014-09-24 20:09:41 by Sir Sir
This was almost unchanged used as effect in Function 256b
added on the 2015-05-19 16:56:40 by HellMood HellMood
rulez added on the 2015-08-30 01:43:56 by T$ T$

submit changes

if this prod is a fake, some info is false or the download link is broken,

do not post about it in the comments, it will get lost.

instead, click here !

[previous edits]

add a comment