high performance computing on graphics processing units: hgpu.org

Fractal image compression provides immense advantages as compared to conventional image compressions. Though the fractal image encoding time is comparatively quite high as compared to the conventional ones but the decoding time is far less and almost instantaneous. Besides, fractal images are resolution-independent, implying that these images will render the same intensity and quality even when scaled. In other words the number of pixels remains unchanged even while extrapolating the image. In addition to it, the fractal image quality remains un-altered even at low-bit rates thus making it a suitable candidate for offline applications. The present baseline approach for fractal image compression is modified and supported with advanced parallel hardware in the form of Graphical Processor Units from Nvidia Corporation. The GPUs consist of many cores thus providing SIMD parallel processing capability at an un-imaginable rate of around 24 GFLOPS. This processing speed was not possible earlier before the advent of GPUs except in some selected highly evolved supercomputers. The rendering of image and its compression is implemented using OpenCL library. The benefits of faster fractal compression lie in the realm of medical imaging, satellite reconnaissance, gaming & film media.