Such instructions are used for audio and video encoding and decoding, games and other applications where the same operation needs to be applied to large amount of data Float Shaders (Pixels/s) ??? higher results are better, i.e. better single (32-bit) floating-point performance.