Java-Gaming.org Java4K winners: [ by our judges | by the community ]         
Featured games (67)
games approved by the League of Dukes 		Games in Showcase (∞)
games submitted by our members



News: Read the Java Gaming Resources, or peek at the official Java tutorials
 
    Home     Help   Search   Login   Register   
Java-Gaming.org > Game Development > Performance Tuning > using JNI math libraries
Pages: [1]
  Print  
  using JNI math libraries  (Read 1455 times)
0 Members and 2 Guests are viewing this topic.
Offline Aaron_Anderson

JGO n00b
*

Posts: 18


perception is reality
« on: 2005-09-26 19:31:08 »
I was reading on the AMD dev forums and came across the article http://www.nag.co.uk/IndustryArticles/HighPerformanceMathLibraries.pdf and noticed the order of magitude performance gains by using the AMD Core Math Library. I also came across the site https://mtj.dev.java.net/ with Java bindings for BLAS.  Would there be any benefit of using these libraries for heavy matrix manipulations once the overhead of JNI was factored in?
Offline Mark Thornton

Sr. Member
**

Posts: 473
Medals: 2
« Reply #1 on: 2005-09-27 04:07:48 »
Quote from: Aaron_Anderson on 2005-09-26 19:31:08
Would there be any benefit of using these libraries for heavy matrix manipulations once the overhead of JNI was factored in?
Maybe. As it would be undesirable for the JNI to lock large arrays for the duration of some of these algorithms (blocking GC), it would probably be preferable to keep the matrices in direct allocated java.nio.Buffer type objects. This may then make the Java api less friendly. Then depending on the problem size you may be better off with something like the Colt libraries instead (which are pure Java).
http://hoschek.home.cern.ch/hoschek/colt/.

Using JNI would be even worse if the JVM copied data rather than merely pinning the array at a fixed location. I don't know how many JVM do this, but the specification permits it. Again using buffers avoids this, but it isn't cost free.
Offline princec
« League of Dukes »

JGO Kernel
*****

Posts: 8089
Medals: 96


Eh? Who? What? ... Me?
« Reply #2 on: 2005-09-27 05:58:15 »
Quote from: Mark Thornton on 2005-09-27 04:07:48
Quote from: Aaron_Anderson on 2005-09-26 19:31:08
Would there be any benefit of using these libraries for heavy matrix manipulations once the overhead of JNI was factored in?
Maybe. As it would be undesirable for the JNI to lock large arrays for the duration of some of these algorithms (blocking GC), it would probably be preferable to keep the matrices in direct allocated java.nio.Buffer type objects. This may then make the Java api less friendly. Then depending on the problem size you may be better off with something like the Colt libraries instead (which are pure Java).
http://hoschek.home.cern.ch/hoschek/colt/.

Using JNI would be even worse if the JVM copied data rather than merely pinning the array at a fixed location. I don't know how many JVM do this, but the specification permits it. Again using buffers avoids this, but it isn't cost free.

I'd just have to say that it is probably unlikely GC will be occuring at this stage in a game environment - you'll probably have one thread doing all the computation (and hence, all the allocation), and therefore no garbage will be generated while this JNI is going on.

Cas Smiley
Play Revenge of the Titans - Play Titan Attacks - Play Droid Assault
Games published by our own members! Go get 'em!
Offline Mark Thornton

Sr. Member
**

Posts: 473
Medals: 2
« Reply #3 on: 2005-09-27 07:10:28 »
Quote from: princec on 2005-09-27 05:58:15
I'd just have to say that it is probably unlikely GC will be occuring at this stage in a game environment - you'll probably have one thread doing all the computation (and hence, all the allocation), and therefore no garbage will be generated while this JNI is going on.
This would be a factor that someone planning a library for more general use might (should) consider. On the other hand if you do your own JNI for use solely in games then you could ignore the issue.

What sort of a game would require the kind of matrix size that would make using these libraries via JNI worthwhile?
Offline Raghar

Sr. Member
**

Posts: 331


Ue ni taete 'ru hitomi ni kono mi wa dou utsuru
« Reply #4 on: 2005-09-27 15:14:12 »
Quote from: Aaron_Anderson on 2005-09-26 19:31:08
I was reading on the AMD dev forums and came across the article http://www.nag.co.uk/IndustryArticles/HighPerformanceMathLibraries.pdf and noticed the order of magitude performance gains by using the AMD Core Math Library. I also came across the site https://mtj.dev.java.net/ with Java bindings for BLAS.  Would there be any benefit of using these libraries for heavy matrix manipulations once the overhead of JNI was factored in?
What about some benchmark?
Pages: [1]
  Print  
 
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.16 | SMF © 2011, Simple Machines Valid XHTML 1.0! Valid CSS!
Page created in 0.072 seconds with 19 queries.