c906508e62
- separate jar files for engine components - resolve dependencies between code parts - remove Nifty dependency from Cinematics - remove Physics dependency from TerrainGrid - add public accessors to Natives Extraction - remove RenderHint serialization from networking git-svn-id: https://jmonkeyengine.googlecode.com/svn/trunk@8839 75d07b2b-3a1a-0410-a2c5-0572b91ccdca
32 lines
948 B
Plaintext
32 lines
948 B
Plaintext
CPU usage:
|
|
----------
|
|
|
|
ObjLoader.load()
|
|
- ObjLoader.readLine()
|
|
|
|
HDRLoader.writeRGBE() // need faster RGBE8 -> RGB16F conversion
|
|
|
|
// OpenGL resource-intesive points
|
|
Renderer.renderQueue()
|
|
Renderer.setVertexAttrib()
|
|
Material.apply()
|
|
|
|
Memory usage:
|
|
-------------
|
|
- OBJLoader
|
|
Java's Scanner class allocates approx. 8 MB of memory
|
|
to load the teapot model. Either implement ObjLoader without Scanner
|
|
or create an import/export system!
|
|
|
|
- AWTLoader
|
|
Using AWT for loading images is slow and uses more memory
|
|
than a home-grown loader. Use DDS and TGA formats more.
|
|
|
|
- Shader.getUniforms
|
|
This method generates a collection to represent the Uniforms
|
|
in the shader and is used by Renderer.updateShaderUniforms()
|
|
Need a faster method to iterate & update uniforms in a shader.
|
|
|
|
- Material.apply
|
|
Same thing as above. Generates a Collection and then an Iterator for a HashMap.
|
|
First, consider if using a HashMap is neccessary.. |