Just to let everyone know I've downloaded and installed JDK 1.4 beta 3 and the performance issues I posted about below have all disappeared. Yay!
But, unfortunately, a bug has appeared in beta 3 that screws up my test framework. Damon. Damon Maria wrote: > > http://java.sun.com/products/java-media/2D/perf_graphics.html > > Hi all, > > having read thru the above document I'm currently writing a test of JDK 1.4's > Java2D performance for a project. I know it's in beta still but... > > In the section "Tips for Attaining Better Performance" it talks about > acceleration of images with 1-bit transparency (which I assume is > Transparency.BITMASK). To quote: > > "Acceleration support for 1-bit transparency enables you > to accelerate sprites with transparent pixels so..." > > However when I gc.createCompatibleImage( width, height, Transparency.BITMASK ) > the resulting drawImage()'s are really slow. My test (filling the screen with > 10x10 of these images) for different transparencies passed into > createCompatibleImage() results in: > > OPAQUE 25 fps > BITMASK 1 fps > TRANSLUCENT 12 fps > > fps = frames/second -> higher = better > > Strange, I would have thought BITMASK would have been close to OPAQUE because it > would be accelerated by DirectDraw, and TRANSLUCENT would be the slowest. Can > someone explain why BITMASK isn't accelerated as the document says it should be. > > Also, how do I create a VolatileImage with 1-bit transparency? I can't seem to > find anyway to do it? There is no createVolatileImage( width, hieght, > transparency ). And incidentally, using a VolatileImage in the above test > results in 50 fps. > > My system: > Windows 2K, DirectDraw 8, Riva TNT 2, JDK 1.4 beta 2. > > thanks in advance, > Damon. =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA2D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
