iACT: A Software-Hardware Framework for Understanding the Scope of Approximate Computing

Approximate computing has recently emerged as a paradigm for enabling energy efficient software and hardware implementations by exploiting the inherent resiliency in applications to impreciseness in their underlying computations. Much of the previous work in this area has demonstrated the potential for significant energy and performance improvements, but these works largely consist of ad hoc techniques that are applied to a small number of similar applications. Mainstream adoption of approximate computing requires a deeper understanding of the inherent application resilience and the codesign hardware to go with the software. This dictates tools and methods that can help programmers reason about the scope and behavior of approximations in applications. To this end, this paper discusses an open source toolkit, called iACT (Intel’s Approximate Computing Toolkit) to analyze and study the scope of approximations in applications. Our toolkit consists of a compiler, runtime and a simulated hardware test bed. We discuss the design of this toolkit and present examples of how to use this toolkit in this paper. As an example on how to use this toolkit, we include two different applications and analyze the scope of approximate computing in these.