A number of ARM-based application processor vendors are shipping or have announced 3D accelerated devices. These devices span a broad spectrum of capabilities and performance. This article provides an insight into the complexities involved with developing high-performance 3D applications for these platforms. Topics covered will include fixed-point/floating point abstraction for optimal CPU usage, optimisation advice for OpenGL ES, texture management and compression, optimal geometry submission, vertex shader based geometry processing and general tips and tricks.

Reprinted in its entirety from ARM IQ Vol. 4, No. 3, 2005