Benchmark questions

Hi folks,

I'm testing Jme3, in some comparisions i try to load simple objects and see the fps, to compare i'm using the last release of fraps (

I use the object in Jme3 svn: Tests-Data/models/teapot.obj

The code i use in jme3 in NetBeans:

        Geometry teaGeom = (Geometry) manager.loadModel("teapot.obj");
        Material mat = new Material(manager, "debug_normals.j3md");


The code i use in jme2 in Eclipse:

      Spatial player = null;
      String model = "com/tumbu/learning/game/teapot.obj";
      URL path;
      ObjToJme objToJme = new ObjToJme();
      ByteArrayOutputStream BO = new ByteArrayOutputStream();
      try {
         path = Convert.class.getClassLoader().getResource(model);
         player = (Spatial)BinaryImporter.getInstance().load(new ByteArrayInputStream(BO.toByteArray()));
      } catch (IOException e) {

I'm using GeForce 9600 gso, Core 2 Duo E8400, 3gb ram, Windows 7

Jme3: 4200 fps
Jme2: 2800 fps


Uhm, Fraps is a tool to record video from your desktop, isnt it? So the fps it shows is either the refresh rate of your monitor (60Hz or something) or the framerate of the video it records (25fps or something). The framerate that jme3 shows is the actual number of OpenGL render updates per second.

The differences in jme2 and jme3 are due to the fact that both are quite different since jme3 was written from scratch to bring next-generation performance

Hope this helps!



The difference in fps in one teapot rendered doesn't really mean anything. You should render more detailed scenes with more advanced shading, then you can compare the fps.

thanks momoko_fan

maybe you know a test written in test package of jme2 or jme3 to do the comparision?