I've written an obj model loader in opengl es, and the native format is floating point variables. Right now the model is loaded into float buffers and then passed to opengl. I know that these phones don't have floating point processors so this can be detrimental to performance, and I've noticed when loading files over 100kb it can take a few seconds. I want to increase the performance but a little unsure as to what data type would be most efficient, and still be opengl friendly. I know this is simple question but I've always been able to work with floating points. I think i can use a doublebuffer, but is this the most efficient way? Also, when passing this buffer to opengl, I assume ill have to change the type from GL_FLOAT? Any input on the matter would be appreciated

Check out this thread, towards the bottom. I think your problem might actually be the floatbuffers themselves.

android-2d-3d-graphics-opengl-problems-f55/updating-a-vbo-t17606.html

You may want to try one of the mentioned solutions before switching to GL_FIXED for all your stuff.