219 replies to this topic

[synkro]

Or we could just try and get as most effects into as small an exe as possible (I'm thinking 40KB or so...)

Or just full of printfs a full text demo only for people with a serial cable...

[Franxis]

It is very difficult to get an executable < 64 Kb in GP2X. If you use DevKitGP2X you have to use -static flag.

Also it has no sense to don't use executable compression, it is used in almost all 64 kb intros.

[Dzz]

It is very difficult to get an executable < 64 Kb in GP2X.

I would say it is "slightly" difficult. As with any platform, solving these type of problems is part of the fun of writing demos.

[Squidge]

It is very difficult to get an executable < 64 Kb in GP2X. If you use DevKitGP2X you have to use -static flag.

Also it has no sense to don't use executable compression, it is used in almost all 64 kb intros.

Although the smallest executable created by DevKitGP2X and the like seems to be a few hundred KB, it is certainly possible to get much smaller by using various techniques. Perhaps when the demos are released, people will be able to learn some of these techniques

Also, yes, it would be good to use executable compression, but I think at the moment, only Rlyeh has done a compressor/decompressor, but I'm unsure whether his decompressor itself is under 64KB, which would kinda ruin the point of the compression. We'll need a tightly coded decompressor that relies on zero libraries.

[Javacat]

This competition sounds pritty cool. It should (hopefully) produce some really interesting things. I'm hopeing we get some fancy demos using both CPU's and any other chips hidden away.

For anyone wanting to see this sort of demo for the PC you should check out this movie - http://www.albinoblacksheep.com/download/6...ksheep.com).exe. This competition excludes games, but something even more impressive than that movie is @ http://www.theproduk...m/kkrieger.html.

[PofG1]

Amazing, I'm lookigng forward to seeing these.

And sucha big prize

[The Moose]

heres hoping for electroplankton clone!

[synkro]

It is very difficult to get an executable < 64 Kb in GP2X. If you use DevKitGP2X you have to use -static flag.

Also it has no sense to don't use executable compression, it is used in almost all 64 kb intros.

Although the smallest executable created by DevKitGP2X and the like seems to be a few hundred KB, it is certainly possible to get much smaller by using various techniques. Perhaps when the demos are released, people will be able to learn some of these techniques

Also, yes, it would be good to use executable compression, but I think at the moment, only Rlyeh has done a compressor/decompressor, but I'm unsure whether his decompressor itself is under 64KB, which would kinda ruin the point of the compression. We'll need a tightly coded decompressor that relies on zero libraries.

I still don't get rlyeh's gpe compressor to work under Linux but the UCL readme states that the decompressor has a size around 200bytes. I remember that Spiv's on the gp32 packed pretty well...

[Hanz™]

I would be surprised if people did not try and write some demos, this is quite a lot of money. And money talks.

[Orion_]

This compo would almost make me buy a gp2x.

[synkro]

This compo would almost make me buy a gp2x.

Just win and you got a free gp2x, guru status and pussy galore all at once...

[Dzz]

I still don't get rlyeh's gpe compressor to work under Linux but the UCL readme states that the decompressor has a size around 200bytes. I remember that Spiv's on the gp32 packed pretty well...

I have gotten it to work on Windows (well i just downloaded it and ran it) and it works great on large files. The decompression "stub" it puts in the resulting file is about 43kb though, which makes it of marginal use for tiny demos. For the "normal" (non-64kb) demo category there is no real need for a compressor.

I have been intending to make a more appropriate UCL-based compressor as part of my article series; if somebody else beats me to it, that's great. I don't think I'll be working on that for another month or so.

[synkro]

I still don't get rlyeh's gpe compressor to work under Linux but the UCL readme states that the decompressor has a size around 200bytes. I remember that Spiv's on the gp32 packed pretty well...

I have gotten it to work on Windows (well i just downloaded it and ran it) and it works great on large files. The decompression "stub" it puts in the resulting file is about 43kb though, which makes it of marginal use for tiny demos. For the "normal" (non-64kb) demo category there is no real need for a compressor.

I have been intending to make a more appropriate UCL-based compressor as part of my article series; if somebody else beats me to it, that's great. I don't think I'll be working on that for another month or so.

well, I dunno what you want to do with your article series but I am willing to help. (well UCL is abit much for me atm) like putting old, overworked and bug fixed articles into PDFs.. I also have fine fixed point code here (well as we don't have -lm I have a already build a sine LUT)

[aapje89]

Cool cool cool
I cant wait to see some leet demo's!

64kb is definitly possible, look at 64kb pc demos, some have got really really nice 3d effects and are very long.


btw. some very cool pc demos here:
http://asd.demoscene...php?where=prods

[Dzz]

Cool cool cool
I cant wait to see some leet demo's!

64kb is definitly possible, look at 64kb pc demos, some have got really really nice 3d effects and are very long.

We have to walk before we can run so probably the first batch won't be super elite but we'll learn a lot at least.

Those amazing PC demos have several advantages over us: First, they are usually built by teams of 3-10 people who have been optimizing their codebases for a long time. Second, a modern PC is much faster than a gp2x. Third, PC's have very fast floating point performance which helps a lot. Finally (and most importantly) those demos rely on fast hardware-accelerated 3D cards.

Still, the gp2x is also a pretty decent little piece of hardware and we should be able to make it dance at least a little bit!

well, I dunno what you want to do with your article series but I am willing to help. (well UCL is abit much for me atm) like putting old, overworked and bug fixed articles into PDFs.. I also have fine fixed point code here (well as we don't have -lm I have a already build a sine LUT)

Nifty. Don't worry about my article series; I'm not even an expert on demos, I'm just having fun looking into the issues as I try to build one and writing up what I do. So if you want to share some code for building sine tables or doing fixed point math or whatever, I'm sure everybody would appreciate it.

Edited by Dzz, 13 April 2006 - 03:16 PM.