Leaking GL Test problem?

[[

Hi, I have a issue in my project taht causes a outof memory in direct buffer. (that happens when i move around the world much)

However any single point in the world is far under the direct memory limit (around 50 mb needed but at around 800mb it crashes).

Now using the “m” key in simpleapplication I see that direct memory is constantly growing and I don’t know why.

]]

After that I tested the LeakingGLTest and I see the amount of Buffers rising rapidly, however the direct memory stays more or less stable. However I’m kinda worried about the buffer count rising constantly, is this the expected behaviour or is there some problem in the cleaning up of ressources?

Actually the direct memory seems to be leaking slowly as well if I use this slightly modified test

[java]/*

  • Copyright © 2009-2010 jMonkeyEngine
  • All rights reserved.

    *
  • Redistribution and use in source and binary forms, with or without
  • modification, are permitted provided that the following conditions are
  • met:

    *
    • Redistributions of source code must retain the above copyright
  • notice, this list of conditions and the following disclaimer.

    *
    • Redistributions in binary form must reproduce the above copyright
  • notice, this list of conditions and the following disclaimer in the
  • documentation and/or other materials provided with the distribution.

    *
    • Neither the name of ‘jMonkeyEngine’ nor the names of its contributors
  • may be used to endorse or promote products derived from this software
  • without specific prior written permission.

    *
  • THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  • "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
  • TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
  • PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
  • CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
  • EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
  • PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
  • PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
  • LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
  • NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
  • SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

    */

    package jme3test.stress;

    import com.jme3.app.SimpleApplication;

    import com.jme3.material.Material;

    import com.jme3.math.Vector3f;

    import com.jme3.renderer.GLObjectManager;

    import com.jme3.scene.Geometry;

    import com.jme3.scene.Mesh;

    import com.jme3.scene.Node;

    import com.jme3.scene.Spatial.CullHint;

    import com.jme3.scene.shape.Sphere;

    import java.util.logging.Level;

    import java.util.logging.Logger;

    /**
  • Generates 400 new meshes every frame then leaks them.
  • Notice how memory usage stays constant and OpenGL objects
  • are properly destroyed.

    */

    public class TestLeakingGL extends SimpleApplication {

    private Material solidColor;

    private Sphere original;

    public static void main(String[] args){

    TestLeakingGL app = new TestLeakingGL();

    app.start();

    }

    public void simpleInitApp() {

    original = new Sphere(4, 4, 1);

    original.setStatic();

    // original.setInterleaved();

    // this will make sure all spheres are rendered always

    rootNode.setCullHint(CullHint.Never);

    solidColor = (Material) assetManager.loadAsset("Common/Materials/RedColor.j3m");

    cam.setLocation(new Vector3f(0, 5, 0));

    cam.lookAt(Vector3f.ZERO, Vector3f.UNIT_Y);

    Logger.getLogger(Node.class.getName()).setLevel(Level.WARNING);

    Logger.getLogger(GLObjectManager.class.getName()).setLevel(Level.WARNING);

    }

    @Override

    public void simpleUpdate(float tpf){

    rootNode.detachAllChildren();

    Mesh sphMesh = new Sphere(4, 4, 1);

    Geometry sphere = new Geometry("sphere", sphMesh);

    sphere.setMaterial(solidColor);

    sphere.setLocalTranslation(1.5f, 0,1.5f);

    rootNode.attachChild(sphere);

    }

    }

    [/java]

Actuall the more I dig into this problem the more I’m aware of a fundamental JME design mistake,

Currently the Assetstuff seems to use WeakReferences, wich can be GC as soon as they are not any longer nstrong referenced, while a Softreference would be far more intelligent here, as it tends to stay in memory as long as there is enough aviable.

More when i know where the growing buffer count comes from.





(Also if anybody wants to know the differenes, thishere explains it quite good http://weblogs.java.net/blog/2006/05/04/understanding-weak-references)

1 Like

Ok I found the cause of the DirectBuffer out of memory, the ModelKey defines models as cacheable, but not to be cleaned up automatically. In case of much larger meshes(like a city you driving through ) this leads to much allocated DirectMemory that is not needed actually. (I detach stuff to far away,)



When the caching is set to wrong the direct memory stays nearly stable, however the buffercount still rises like hell.