A multi-core reconfigurable architecture for ultra-low power bio-signal analysis

Abstract

Abstract-This paper introduces a novel computing architecture devoted to the ultra-low power analysis of multiple biosignals. Its structure comprises several processors interfaced with a shared acceleration resource, implemented as a Coarse Grained Reconfigurable Array (CGRA). The CGRA supports the efficient execution of the computationally intensive kernels present in this application domain, while requiring a low reconfiguration overhead. The run-time behavior of the resulting heterogeneous system is orchestrated by a light-weight hardware mechanism, which concurrently synchronizes processors and regulates access to the reconfigurable accelerator. The architecture achieves speed-ups of up to 11x on different bio-signal processing kernels and system-level energy savings of up to 18.6%, with respect to a multi-core platform, which does not feature CGRA acceleration.

i-DPs CGRA: An Interleaved-Datapaths Reconfigurable Accelerator for Embedded Bio-Signal Processing

“…A similar conclusion is drawn in [4] and [5], which propose an embedded platform for bio-signal processing in personal health monitors, an increasingly relevant domain with ultra-low power constraints [8]. Similarly to us, [5] describes a CGRA with multi-DP RCs operating in SIMD (single instruction-multiple data) mode.…”

“…7 compares the energy consumption of the part of the workload that is accelerated for the two considered benchmarks. Again, three architectural choices are considered: a multi-processor platform [2], that does not embed a reconfigurable mesh; a platform that couples a multi-core platform with a single-DP CGRA [4]; and our proposed system, featuring an i-DPs accelerator. It can be noted that, even in the two latter cases, certain software overhead is required for setting up, launching and retrieving the outputs of an acceleration request.…”

HEAL-WEAR: An Ultra-Low Power Heterogeneous System for Bio-Signal Analysis

“…Conversely, we instead concurrently support requests by different cores, either in a Multiple-Instruction / Multiple-Data (MIMD) or in a SIMD fashion. As opposed to [46], which only considers SIMD at the multi-processor level, our paper investigates the benefits of also supporting SIMD-kernels, adopting a multi-datapath CGRA.…”

“…Hence, the execution of SIMD kernels is not supported, and SIMD acceleration requests are mapped on different CGRA regions. This setup corresponds to the platform described in [46].…”

A Survey on Coarse-Grained Reconfigurable Architectures From a Performance Perspective

“…The CGRA they constructed had the additional benefit of being able to powerdown sections of the CGRA when unused to extend battery life. Another biomedical CGRA was introduced by Duch et al [158], and uses a mesh-like composition and a 1 MHz clock-frequency to accelerate electrocardiogram (ECG) analysis kernels.…”