Mailing Lists: Apple Mailing Lists

Image of Mac OS face in stamp
 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Fullscreen test sample code



On Mar 17, 2004, at 7:25 PM, Lynn M. Meredith wrote:

Odin

I compiled your source example with Project Builder v 2.1. Here is my config:

+ PowerBook G4 800MHz
+ 256 KByte L2 Cache
+ 1 MByte L3 Cache
+ 512 MByte RAM
+ MacOS 10.3.3
+ Java Version 1.4.2_03 (I have not upgraded to Developer Preview 1 of Java 1.4.2 Update 1)
+ ATI Rage M7, 32 MByte VRAM
+ Page Flipping = false
+ Backbuffer accelerated = true

I get 28 FPS.

Regards, Lynn

On Mar 17, 2004, at 11:19 AM, Odin Jensen wrote:

Hi guys.

I threw together a simple fullscreen scroll tile thingie for the test.
It runs at 20 fps on my system (IBook G3 900mhz, Radeon 8500 32 meg, 256 meg ram OS X newest patch, java 1.4.2 newest developer update.)

Can anyone see if I'm doing anything wrong here. This sample gets hardware acceleration on Windows.

Grab it at http://homepage.mac.com/ozak

Thanks in advance

Odin

It is interesting that running in it in a window instead of full screen I get around 108 FPS while when running full screen I get 33 FPS. If I use the window's default buffer strategy (comment out the call to createBufferStrategy) I get around 40 FPS in a window and full screen I get 36 FPS. I also note no difference between using a buffered image and a volatile image.

This is on a PM Dual G5 2Ghz / ATI 9800 128 VRAM / Mac OS 10.3.3 / Java 1.4.2_03 while the system is loaded doing other things.

Anyway see the following email from an Apple guys a little while back on this subject...

-Shawn


On Dec 21, 2003, at 5:54 PM, gerard ziemski wrote:

John,

On Dec 19, 2003, at 6:18 PM, email@hidden wrote:

Message: 10
Date: Fri, 19 Dec 2003 16:34:46 -0800
Subject: Re: Java 1.4.1's lack of hardware acceleration MUST be defined
From: John Osborne
To: email@hidden

Gerard,

Thanks for your response. I'm curious why hardware acceleration was
present in 1.3.1 and removed in 1.4.x? Was it a hack? Was HWA removed
from core graphics?

1.3.x used OpenGL to do its hwaccel. CoreGraphics is a 2D graphics API (currently all done in software), OpenGL is a 3D graphics API (hardware based). The hwaccel mechanism in 1.3.x was a layer we wrote, that sits on top of OpenGL and did 2D stuff.

The hwaccel in 1.3.x was not removed from 1.4.x, it is just 1.4.x is a totally new rewrite (from Carbon to Cocoa), and it was never ported over.


I'm also curious whether anyone on this list knows
whether the Window's JVMs benefit from hardware acceleration? I'm also
curious why the graphics performance of Java3D seems on par with the
Windows equivalent when Java2D is quite a bit slower?

Java3D is a set of 3D graphics APIs, so it is straightforward to implement on top of OpenGL, hence better speed. Trying to use OpenGL to do 2D stuff is much more difficult, but doable, as we showed in 1.3.x.

Instead of porting the 1.3.x hwaccel mechanism to 1.4.x we decided to spent the time working on the quality of our 2D implementation. Once CoreGraphics is done with QuartzExtreme2D, or Sun includes their own OpenGL based graphics pipelines, we'll have hwaccel graphics in JDK on Mac OS X.


cheers
_______________________________________________
java-dev mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/java-dev
Do not post admin requests to the list. They will be ignored.


References: 
 >Fullscreen test sample code (From: Odin Jensen <email@hidden>)
 >Re: Fullscreen test sample code (From: "Lynn M. Meredith" <email@hidden>)



Visit the Apple Store online or at retail locations.
1-800-MY-APPLE

Contact Apple | Terms of Use | Privacy Policy

Copyright © 2007 Apple Inc. All rights reserved.