Welcome, Guest
Want to take part in these discussions? Sign in if you have an account, or apply for one below
-
- CommentAuthorfxkill
- CommentTimeNov 29th 2011
I've been working on a snowglobe app for the iphone, but I'm having trouble with the refresh rate with only 100 particles; it drops to around 10fps. I've been using the internal flint 2D bitmapRenderer with two shared images, and I've tried using both CPU and GPU render modes with little effect. Is there a different method of rendering I should be using? -
- CommentAuthorFrinkky
- CommentTimeDec 3rd 2011
I found the DisplayObjectRenderer better performance. The bitmapRenderer has to rasterise eash frame into a bitmap which more than likely adds to the processing overhead. Are you using RadialDot's or your own particle object? -
- CommentAuthorRichard
- CommentTimeDec 14th 2011 edited
This is very much a process of trying everything. cpu vs gpu mode, different renderer types, and different particle objects.
Depending on your frame-rate, and your target frame-rate, you can also make Flint run less frequently than every frame, to free resources for the rest of your app. To do this, set the useInternalTick property of the emitter to false, and set up your own timing mechanism. Then call the emitter's update method based on your timing mechanism, passing it the time in seconds since the last update.
There are some inefficiencies in Flint, most of which are the result of its flexibility. When I created this playbook project - http://www.richardlord.net/particle-soup - I used Flint until I had what I needed, then I removed bits of Flint that I wasn't using (e.g. checking whether particles should be removed - in this instance that wasn't necessary) to make it more efficient. You may find that to get the performance you need, you need to remove unwanted bits of Flint.
You may even find the only solution is to hand code your particle effect. But Flint can still be used to figure out exactly what you want so you only have to hand code it once, and to provide the algorithms for the effects.
1 to 3 of 3
