Reduction in Data

Software developed for the ECM353X family, automatically and transparently optimizes code execution across asynchronous cores for easy-to-use application implementation. Specially developed algorithms take advantage of DSP architecture doing two multiplies and accumulates (MACs), with dual memory banks and zero loop overhead with address generation. For example, when compared to an M4F running CMSIS DSP matrix math functions, the ECM3532 performs 2.5X more matrix multiplies.