Higher Computational Throughput With a Slower Clock: MCU Architectures for Compute-Intensive Embedded Applications
The increasing use of complex algorithms in embedded control systems is substantially adding to processing overhead. Fast Fourier Transforms (FFTs), inverse Discrete Cosine Transformation (iDCTs) and other compute-intensive algorithms that require single-bit manipulation, matrix mapping in addition to byte and half-word arithmetic are becoming common in applications that were unimaginable just a few years ago. These include data compression, signal coding, vehicle passenger safety systems, and portable infotainment systems which use MP3 audio and MPEG-4 video codecs. All require advanced, computationally-intensive DSP algorithms.
This paper introduces Atmel's AVR32 core, which exhibits superior instruction execution throughput, code density and power consumption for cost-sensitive, power-constrained, DSP-based embedded applications.