[QUOTE=Richy19;31801794]Ups, I meant uniform and attribute variables in the shaders[/QUOTE]
Uniforms are the same for all vertices in the mesh you're drawing, attributes are different for every vertex. There are also varying variables that are interpolated for every triangle meaning that every fragment (pixel) has it's own unique variable.
You can think of is this way:
You're rendering 10 objects, each of those object has it's own set of uniforms.
Inside one of those objects, each vertex has it's own attributes.
And finally, when the triangles are rendered, each of the pixels between 3 vertices as it's own varying variables.
[media]http://www.youtube.com/watch?v=uK6vNTi5c5s[/media]
Sokoban! Haven't done much other than box pushan, though. Might experiment with procedurally generated puzzles later or something.
Goddamnit. I guess XNA does not support SIMD? I have a buttload of particles that I need to render/update and I want a really fast way to do it without having to rewrite my entire particle system to be GPU based...
[QUOTE=BlkDucky;31801936]-video-
Sokoban! Haven't done much other than box pushan, though. Might experiment with procedurally generated puzzles later or something.[/QUOTE]
Ahh, I'd play that all my day...
working on a full 2D engine with OpenGL (C++), as I don't want to use any libraries (because I love doing all the small things myself) I started with the most basic thing which is bitmap fonts, I know it's not a huge achievement but hey, I like it ;p
[IMG]http://www.binarybit.ch/imgs/bitmap.jpg[/IMG]
next up: Menu, particle engine, physics (verlet)!
PS: top right FPS is Fraps (it's limited to 100 fps)
[media]http://www.youtube.com/watch?v=DvafigrwiaM[/media]
2 videos in one page, yeah.
[QUOTE=Foda;31802683]Goddamnit. I guess XNA does not support SIMD? I have a buttload of particles that I need to render/update and I want a really fast way to do it without having to rewrite my entire particle system to be GPU based...[/QUOTE]
Have a look at the mercury library.
I cant promise anything, but as far as I know it's pretty fast.
Well, my particle-system can produce 250 particles each frame, I don't know if that's enough, but it's alot of particles!
(Each particle has a different color, angle, spinningpower, size and lifetime.)
All without lag.
[QUOTE=Staneh;31805618]Well, my particle-system can produce 250 particles each frame, I don't know if that's enough, but it's alot of particles!
(Each particle has a different color, angle, spinningpower, size and lifetime.)
All without lag.[/QUOTE]
If it can run with 250 particles without lag, what prevents it from operating with more?
[QUOTE=Deco Da Man;31805689]If it can run with 250 particles without lag, what prevents it from operating with more?[/QUOTE]
He said 250 particles [i]each frame[/i]. Which means...he's Euclideon and he has infinite particles :v:
Well not exactly unlimited, but, at like 300 particles e/f it begins to lag slightly, at 500 e/f it's really fucking laggy and at like 600 e/f or above it just freezes when it has enough particles. But, here, a pic of 250 e/f:
[img]http://i54.tinypic.com/30t6yat.jpg[/img]
[img]http://i56.tinypic.com/aet8uo.jpg[/img]
XNA can take the particles
[media]http://www.youtube.com/watch?v=CyAZ2Y7nOTw[/media]
[QUOTE=Night-Eagle;31805716]He said 250 particles [i]each frame[/i]. Which means...he's Euclideon and he has infinite particles :v:[/QUOTE]
let's ask notch for his professional opinion
Oh my god, well that's alot better then mine.
[QUOTE=ROBO_DONUT;31799517]It's a single dot-product and a scalar subtraction...
It's seriously like four instructions without SIMD.[/QUOTE]
I am doing point-line segment distance which is a bit more complicated.
Well heres the implementation I found online:
[cpp]
// Return minimum distance between line segment vw and point p
const float l2 = LengthSquared(One - Two);
if (l2 == 0.0) return Distance(Point, One);
const float t = VecDot(Point - One, Two - One) / l2;
if (t < 0.0) return Distance(Point, One); // Beyond the 'v' end of the segment
else if (t > 1.0) return Distance(Point, Two); // Beyond the 'w' end of the segment
const sf::Vector2f projection = One + t * (Two - One); // Projection falls on the segment
return LengthSquared(Point-projection);
[/cpp]
I recently sped it up a bit by getting rid of the square root seeing as I am checking if the distance is close to 0 and getting the true distance is not really needed.
Anyway that's besides the point, a separating axis test is more suited to collision detection because it can test how much the shapes are penetrating which means I can adjust impulses and find how to fix penetration more easily than with my current method.
[QUOTE=BlkDucky;31805244][media]http://www.youtube.com/watch?v=DvafigrwiaM[/media]
2 videos in one page, yeah.[/QUOTE]
Meanwhile, at the Legion of Doom
[img]http://i.imgur.com/U7A6C.png[/img]
yay edge detection
Hey guys! My app, Helios, has been released in the App store for about a week now and I forgot to post about it here!
[url]http://itunes.apple.com/us/app/helios-bluetooth-chat/id455896218?ls=1&mt=8[/url]
Helios currently works on iOS 4.3 or greater. I am working to add 4.2 support for version 1.1.
It would be really cool if you could get it and show it to your friends. I am taking feedback here and at [url=mailto:jerish@i3software.org]my email[/url], so if you have a request for a feature just send it my way.
[quote]
Rated 12+ for the following:
* Infrequent/Mild Profanity or Crude Humor
* Infrequent/Mild Alcohol, Tobacco, or Drug Use or References
[/quote]
Where is any of that found in your app? Are they rating potential user interactions?
[QUOTE=Irtimid;31807793]Where is any of that found in your app? Are they rating potential user interactions?[/QUOTE]
I had to put that there. Its a legal thing, but yeah. Potential user interactions.
So when you install the app, there is none of that. Its in the chat where that rating comes into play.
[img]http://eagle.undo.it:8083/img/kintel_62.png[/img]
Okay, yeah, pretty sure that's not supposed to happen - but I can work with this...
[QUOTE=Richy19;31805829]XNA can take the particles[/QUOTE]
[IMG]http://i.imgur.com/A6QPp.png[/IMG]
Work PC:
P4 @ 3.0ghz, onboard video, so I suppose the performace is not too bad
[QUOTE=Foda;31808581][IMG]http://i.imgur.com/A6QPp.png[/IMG]
Work PC:
P4 @ 3.0ghz, onboard video, so I suppose the performace is not too bad[/QUOTE]
Your game looks beautyful
[QUOTE=Night-Eagle;31808310][img]http://eagle.undo.it:8083/img/kintel_62.png[/img]
Okay, yeah, pretty sure that's not supposed to happen - but I can work with this...[/QUOTE]
You accidentally a whole voronoi diagram?
[QUOTE=ROBO_DONUT;31808649]You accidentally a whole voronoi diagram?[/QUOTE]
[img]http://eagle.undo.it:8083/img/kintel_63.png[/img]
...it's just that there [i]might[/i] be some kind of problem here, and I can't [b][i]quite[/i][/b] put my finger on it...
What the hell could it be!? :smithicide:
Edit:
Seriously, it's vexing...
[code]
Error at site ordering check:
(2.84891, 85.4171)
128(1.88855, 81.7683), 132(-4.78254, 76.6258)
132(-4.78254, 76.6258), 108(-49.8514, 71.2473)
108(-49.8514, 71.2473), 20(-635.803, 85.4171)
20(-635.803, 85.4171), 24(-12224.5, 365.661)
304(-7.03702, 99.0954), 308(3.94246, 93.4838)
308(3.94246, 93.4838), 154(4.28017, 85.539)
154(4.28017, 85.539), 158(4.35948, 83.673)
158(4.35948, 83.673), 128(1.88855, 81.7683)
4300(-520298, 32341.9), 83(-235.84, 115.113)
83(-235.84, 115.113), 309(-11.1323, 101.188)
309(-11.1323, 101.188), 304(-7.03702, 99.0954)
24(-12224.5, 365.661), 4301(-3.0511e+06, 78166.5)
[/code]
That's supposed to be the ordered edges of a polygon. Each line is two vertices prefixed by index in CCW (if not, CW) order.
Debugging is fun.
Edit:
Okay, yeah, I've been staring at the exact problem for the last 11 minutes and ignoring it. Look at the coords above, and you'll see the chain is valid if you start from index 4300. You'll end at 4301. How did I miss that?
Edit:
I did it!
[img]http://eagle.undo.it:8083/img/kintel_64.png[/img]
[img]http://i3software.org/s/SS-2011-08-18_15.26.07.png[/img]
Working on the Repair Shop dialog! What do you guys think? ($100 for 10 health points)
(By the way, this is a mockup I am not done programming the thing)
menu done (up down keys to navigate) works perfectly ;p
[IMG]http://www.binarybit.ch/imgs/menuBAM.jpg[/IMG]
particle engine I'm coming!!!!
[media]http://www.youtube.com/watch?v=8-W_iFDQ9SQ[/media]
Am I enjoying JNA too much?
Decided to write an evolution simulator along the lines of darwin bots because i have nothing better to do. Im making the bot "language" binary though, because it saves a lot of hassle writing a parser for human readable code, plus the aim of this isnt to write my own bots but to see what a large simulation of robots will produce. I dont have anything to show as of yet, im still nailing down some ideas but hopefully in a few days the little buggers will be beating the crap out of each other
Sorry, you need to Log In to post a reply to this thread.