Kelvin Chung wrote: >>Alot will depend on what else you are doing in the Java3D system. In >>the archives are details about the multi-threaded nature of Java3D. >>Depending on your use of sound, behaviors, collision detection you will >>see different responses to multiprocessor use. I think two processors >>will always be helpful, I think it maxed out at about 4... which is the >>most you could really use with most X86 architectures any way. >> > >No, it will not max out at 4. There are many >threads in Java3D, the number depends on >View (x2), VirtualUniverse (x6), >Screen (x1), PhysicalEnvironment (x1), >InputDeviceScheduler (x1) plus 3 others. Most >of them are running in parallel if enough CPUs. > Good to hear and thanks for reposting the threading details. You wouldn't happen to have a nice threading diagram sitting around that we could post on J3d.org would ya?
Part of my analysis is related to the Intel architectures limitations with scaling beyond 4 processors. If you are running Java3D on Sun's hardware then you might see benefit from having more then 4 processor, using X86 architectures I doubt you would, but I've never had a 4 way X86 box so I can't respond from experience. -- Alan Hudson President: Yumetech, Inc. http://www.yumetech.com/ Web3D Open Source Chair http://www.web3d.org/TaskGroups/source/ =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA3D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
