ABSTRACT

The appearance of GPGPUs has changed this situation. Graphics hardware providers have recognized the demand of researchers and developers for using GPUs as a general purpose, highly parallelized computing unit. With the new type of hardware, new application programming interfaces (APIs) have appeared, allowing GPUs to be programmed for non-graphical tasks too. (Liu et al. 2007) All these changes have opened the door to the fully GPU-based, thus highly parallelized material removal simulation of free-form objects. (Tukora & Szalay 2008a) In this paper a simulation method is introduced, where all the simulation tasks are maintained by the graphics hardware, including data creation, representation and manipulation, resulting significant increase of simulation speed and quality.