Power/Performance Exploration of Single-core and Multi-core Processor Approaches for Biomedical Signal Processing

Abstract

Abstract. This study presents a single-core and a multi-core processor architecture for health monitoring systems where slow biosignal events and highly parallel computations exist. The single-core architecture is composed of a processing core (PC), an instruction memory (IM) and a data memory (DM), while the multi-core architecture consists of PCs, individual IMs for each core, a shared DM and an interconnection crossbar between the cores and the DM. These architectures are compared with respect to power vs performance trade-offs for a multi-lead electrocardiogram signal conditioning application exploiting near threshold computing. The results show that the multi-core solution consumes 66% less power for high computation requirements (50.1 MOps/s), whereas 10.4% more power for low computation needs (681 kOps/s).

Multi-core architecture design for ultra-low-power wearable health monitoring systems

“…In fact, multi-lead biological signals analysis are often needed to obtain an accurate view of biological events. However, the analysis of these multi-lead signals entails considerably parallel computation opportunities which can be exploited on multi-core processing platforms [9].…”

“…The multi-core architecture proposed in [9], herein referred to as mc-ref architecture, is a starting point of our investigation to explore an energy-efficient multi-core architecture for biomedical applications. To this end, the proposed architecture includes memory management units (MMUs) (cf.…”

“…4) Power/performance trade-offs of the proposed architecture are explored for different target workloads. The results show that the proposed multi-core solution achieves 39.5% and 38.8% power savings with respect to the state-of-the-art [9] at high workloads (637 MOps/s) and low workloads (5 kOps/s) respectively, due to the combination of multiple power management options.…”

“…Parallel computing using multiple cores can alleviate this issue, provided that the algorithms to be executed can be parallelized. To this end, the work presented in [9] explored the power/performance tradeoffs between sequential and parallel near-threshold computations for various biomedical signal processing requirements. The comparison shows that multicore systems do not only solve the performance degradation problem, but also achieve good energy efficiency.…”

“…Also, Yu et al [11] introduced a sub/near threshold processor specialized for low-energy mobile image processing using architecture-level parallelism to compensate the performance loss. The multi-core architectures in [9] and [10] are built for a more general use, however they achieve limited energy efficiency, notably at low workloads.…”