High-Level Synthesis for Approximate Computing: A Survey

The approximate computing paradigm allows designers to create efficient hardware and software designs by leveraging error-tolerance applications such as signal and multimedia processing, computer vision, and machine learning. A recent focus on incorporating approximate computing techniques in high-level synthesis (HLS) provides a new perspective on hardware design by facilitating trade-offs among performance, energy efficiency, and accuracy. However, in the field of HLS for approximate computing (AHLS), there is a lack of systematic review articles and in-depth analyses of the latest methodologies. This paper presents a comprehensive summary of the latest AHLS technologies, while focusing on error estimation, approximation techniques, and design space exploration (DSE). In addition, an analysis of the current research gaps in AHLS is presented. This work aims to provide researchers, engineers, and scholars with a comprehensive theoretical and practical AHLS framework, fostering academic exchange and technological innovation in this study field.