| |
Release: Zodiakos Particle System
Mythril
|
I implemented Omni's suggestion, it halved the framerate on my computer from about 60 to about 30.
Not sure if I need to show you the code, as I didn't really change all that much, just initiated a temporary buffer, filled it with black and lucency of 50% in the beginning of DrawParticles(), changed the drawing functions to draw to this temporary buffer, and blitted that buffer with AdditiveBlit.
It works, but as I said, quite a bit slower than doing particles. However, I'm thinking AdditiveBlit with premade graphics may be what you'd like to try instead of CircleFill.
Tested it out, works superneat! Try making a gradient sphere,and use it with additiveblit on your current code, instead of circlefill! (Also, have setlucent to different values for each part of one particle, so it kind of fades out!)
Posted on 2004-05-17 01:30:17
|
blues_zodiakos
|
Hmm, I'll have to add that, I know just what you mean.
In other news, after looking at the code for two minutes, I implemented everything I was wanting. I improved the amount of maxparticles displayed by about 2 times what they were by implementing the temp buffer. Particles are actually only drawn about 3 times per second now, or whatever ZPS_SYSTEMSPEED equals. In addition, there is also motion blur. :D It's extremely slow, but there's nothing I can do about it really, my technique is about as fast as they come without assembly language.
On a related note, for some reason, using grabregion to copy one picture to another of equal size is about twice as fast as just blitting it. Something odd I thought I'd share.
I imaging once I add those additiveblit spheres you are talking about, I'll upload it tomorrow. :D
Posted on 2004-05-17 05:06:48
|
blues_zodiakos
|
Fourth, and beta release.
It's pretty much complete. Added both the additive particles and the motion blur (as an option, but it's deadly performance-wise, as you said. :D) Gave you both credit in the readme. The additive particles are spiffy.
And I don't really know what else to do with it, so I'm going to leave it be for now if there aren't any huge bugs found with it. I'm probably going to work on my new project, a remake of an 1984 Apple II/Coco 2 game called Robot Odyssey. Yeesh, that game rocked.
Posted on 2004-05-17 21:03:31
|
Mythril
|
:D
Yeah, work on games instead! ;)
Posted on 2004-05-19 11:05:37
|
Zip
|
Just downloaded new version, it's very nice. My one suggestion would be to start falling particles with terminal velocity to save accelerating them. Or you could start them negative like grue said earlier.
Zip
Posted on 2004-05-25 02:54:51
|
RageCage
|
for some reason the newest version wont run on my computer
Posted on 2004-05-25 03:23:12
|
blues_zodiakos
|
What does it do, exactly?
Posted on 2004-05-25 10:46:12
|
Mythril
|
Quote: Originally posted by Zip
Just downloaded new version, it's very nice. My one suggestion would be to start falling particles with terminal velocity to save accelerating them. Or you could start them negative like grue said earlier.
Zip
I believe the reason that some particles look like they start without speed, is that the engine does something like this:
if (y<0) y=0;
I forgot to suggest him to remove it.
Posted on 2004-05-25 11:04:11 (last edited on 2004-05-27 10:47:16)
|
RageCage
|
it runs, it plays music, but doesnt display anythign on the screen, I mean nothing, it never finishesrendering the window so all I can see is the top bar. I have to use ctrlaltdel to get out of it. This is when I change the window mode so that I dont have to reset my computer when I run it. =p
Posted on 2004-05-25 13:47:22
|
blues_zodiakos
|
That's really, really odd. Are you sure that you unzipped the archive with path information?
Another tip: if verge ever locks up your computer with an infinite loop or something, you can sometimes escape it by putting your computer to sleep, waking it up again, and right clicking verge 3 in the taskbar and selecting close. Very dirty, but better than a reset. Some computer keyboards have a sleep button on them, and some computers are sleepable by hitting the power button for less than 5 seconds.
Posted on 2004-05-26 16:43:54
|
RageCage
|
I unzipped it and ran verge.exe... this has been tried on my own computer and school computers
Posted on 2004-05-27 00:50:51
|
Overkill
|
the only way to make it work seems to be if you reduce the amount of particles and emitters to a very low amount, which show a lacking, and yet still slow particle effect. I mean, it'd lag with only one rain particle on the screen, sadly...
Posted on 2004-05-27 03:15:41
|
blues_zodiakos
|
That makes absolutely no sense, I'm running it on my computer right now and the maxparticlecount is up in the thousands for some of the particle types.
What's more, I just downloaded the ZPS4.zip file fresh off the file section here, unzipped it to my brothers computer, ran it, and it worked perfectly fine (even better than my computer, actually). The particles were abundant.
The symptoms you are showing sounded exactly as if you did not unzip the archive with path info, but, when I just tried renaming that folder to test that hypothesis, it simply refused to run. So I have no idea what the problem could be. To reiterate, the fourth demo here runs fine on both computers I've tested, both AMD Athlon XPs(mine a 2400+, his a 2600+) with VIA motherboard chipsets. I can't imagine it should run any differently on any other computer barring an error in the verge executable, other than speed differences. I imagine it should run at full speed now on any computer because of the autolimiting code (unless your computer can't run Verge3 at full speed).
If it truly is a problem with the limiting code, there is a way to test that. Open up System.VC and uncomment the two HookButton statements at the top. Now, during the demo, you can press the ALT button to turn the limiter on and off. If it works with one and not the other, there's obviously a problem with my limiting code. It's so simple, though, that I doubt it.
Posted on 2004-05-27 12:33:27
|
RageCage
|
I'll play around with this a bit to see if I cant get it working, I've never had any problems with v3 on any of my computers before...
...
Ok well I ran it on another computer, it has new been tested on my winXP 1ghz and 2 win2k computers, one with shitty specs, one with acceptable =p
Its failing in the same way on them all.
after doing a little debugging, it seems to be getting hung up right when showPage is called in the first while loop... I'm gonna try to see if I can figure any more out... oh and just a note, the previous versions do work on my computer.
...
well lemme ask you this, should v3.log have a returned:0 in it? it seems to mean that there's no emitter?
Posted on 2004-05-27 14:51:50 (last edited on 2004-05-27 15:13:16)
|
Buckermann
|
Zodiakos,
Release 4 wouldn't run on my computer too untill I changed (in ZPS.vc)
#define MAXNUMPARTICLES 2000
to
#define MAXNUMPARTICLES 1000
Maybe it's because of my sucky computer :)
Posted on 2004-05-27 15:53:11
|
RageCage
|
hey its workin now, thanx ^_^
Posted on 2004-05-27 16:08:22
|
blues_zodiakos
|
Nah, the returned:0 thing is a holdover from a previous version. Ignore it. :D
It shouldn't matter at all what you put in MAXPARTICLES because the autolimiter changes it dynamically - it's not a define, it's an integer. Within a few seconds of starting the program, it should be set at a comfortable level for your computer. Maybe verge does naughty things with for loops on slower computers, or maybe I made some mistake I'm not aware of. Setting MAXPARTICLES to 1000 is okay though, and I thought I had actually already changed it to that, incidentally. I think it's probably better to change the framerate variables though, if you have a slower computer. It's detailed in the readme. When it seemed like it was frozen, was the debug info being displayed in the corner? If the autolimiter was working correctly, MAXPARTICLES should have slowly decreased until it leveled off.
- edit
Also, people, I really appreciate the feedback. :D If this is working erratically on other computers, I really need to know so I can fix it. Anything you think my contributing to the problems, let me know.
Posted on 2004-05-27 18:12:43 (last edited on 2004-05-27 18:14:57)
|
RageCage
|
I dont think it has to do with 'slow computers'. I tried it on a 2ghz school computer which I think is plenty fast.
Posted on 2004-05-27 19:14:49
|
Buckermann
|
Quote:Originally posted by Blues Zodiakos
When it seemed like it was frozen, was the debug info being displayed in the corner? If the autolimiter was working correctly, MAXPARTICLES should have slowly decreased until it leveled off.
Not for me. In fullscreen I had only a black screen with the music playing, and in window mode I had only a "empty" window. With empty I mean the same effect you get when you just put
while(1) { }
into you autoexec while in window mode.
For me it looked like verge was trapped in a endless loop right after starting the music.
Posted on 2004-05-27 19:16:18 (last edited on 2004-05-27 19:17:11)
|
Buckermann
|
If I understand it correctly you assign MAXNUMPARTICLES to NUMPARTICLES at the start. This is then used in ProcessParticles() and DrawParticles(), timer hooked via zps_timer(), to get a good paricle count for the computer.
Is it possible, that ProcessParticles() and/or DrawParticles() needs so much time for the initial run (on slower PCs) with 2000 particles that zps_timer is again triggered (by the timer event) before the two functions are finished?
As I understand it, this would indeed lead to a endless loop, every time calling ProcessParticles()/DrawParticles() again with the starting value of MAXNUMPARTICLES.
Posted on 2004-05-27 19:35:15
|
|
|
