Abstract Reducing Memory References for FFT Calculation

Fast Fourier Transform (FFT) is one of the most widely used algorithms in digital signal processing. It is used in many signal processing and communication applications. many of the FFT operations are performed in embedded systems. Since Embedded systems is very small processors used in almost all type of appliances from microwave ovens to cars, and many embedded systems are portable and depend on small batteries for power; low energy design is extremely important in embedded systems design. One of the major energy consumption sources in any processor is memory access. Memory access requires more energy than almost any operation in a DSP (Digital Signal Processor), or embedded processor, reducing memory access plays a very important role in reducing energy consumption. In this paper we concentrate on the energy consumption in memory in calculating FFT. we compare between three different techniques in calculating FFT with reference to energy consumption in memory access. We also investigate the effect of the number of registers in the CPU on reducing energy consumption in memory access. 1