I ran both versions of gears.java of both libraries, and JoGL ran much much faster, even on a larger window.
so… is it faster than?
first of all, both libs are basically the same speed. They both use OpenGL and in any decent game the GPU should be the limiting factor (see other thread regarding bindings and benchmark).
That said, my general rule is that lwjgl is faster in its native window than jogl since it doesn’t have to play along with awt. My own benchmarks show that lwjgl in its native window is a lot faster than in awt.
After turning off vsync (got 85 fps in jogl gears :D) I got these numbers:
[quote]GL_VENDOR: NVIDIA Corporation
GL_RENDERER: GeForce 6600 GT/AGP/SSE2
GL_VERSION: 2.0.1
[/quote]
on a 2.4@2.7 GHz p4 (yes I know, I need some core 2 duo action) (jdk 1.6)
lwjgl gears, native display: ~ 3000 fps (~3500 without fraps ???)
lwjgl gears, awt: ~1500 fps (1700 without fraps)
jogl gears, awt (from here: http://download.java.net/media/jogl/builds/archive/jsr-231-webstart-current/Gears.jnlp): 1300 fps (no idea what it is without fraps)
Now bear in mind that there might be a small window size difference or code difference or whatever, I would say that lwjgl and jogl in awt mode is somewhat similar in speed (enough to stop comparing). However I think that jogl is much more stable with regards to multi monitor setups.
from my pov, jogl is more of a generic opengl binding whereas lwjgl is much more tilted towards creating games. Both api’s are ofcourse usable on similar types of apps, its just that the focus isn’t the same.