On Memory Accelerated Signal Processing within Software Defined Radios

Abstract: Since J. Mitola's work in 1992, Software Defined Radios (SDRs) have been quite a hot topic in wireless systems research. Though many notable achievements were reported in the field, the scarcity of computational power on general purpose CPUs has always constrained their wide adoption in production environments. If conveniently applied within an SDR context, classical concepts known in computer science as space/time tradeoffs can be extremely helpful when trying to mitigate this problem. Inspired by and building on those concepts, this paper presents a novel SDR implementation technique which we call Memory Acceleration (MA) that makes extensive use of the memory resources available on a general purpose computing system, in order to accelerate signal computation. MA can provide substantial acceleration factors when applied to conventional SDRs without reducing their peculiar flexibility. As a practical proof of this, an example of MA applied in the real world to the ETSI DVB-T Viterbi decoder is provided. Actually MA is shown able to provide, when applied to such Viterbi decoder, an acceleration factor of 10.4x, with no impact on error correction performances of the decoder and by making no use of any other typical performance enhancement techniques such as low level (Assembler) programming or parallel computation, which though remain compatible with MA. Opportunity for extending the MA approach to the entire radio system, thus implementing what we call a Memory-Based Software Defined Radio (MB-SDR) is finally considered and discussed.