Capri: A Control System for Approximate Programs

TL;DR

Capri is a control system that uses machine learning to optimize approximation levels in programs, balancing resource usage and accuracy through a constrained optimization approach.

Contribution

This work introduces Capri, a novel system that controls approximation in non-streaming programs by modeling and optimizing resource-accuracy trade-offs using machine learning.

Findings

Effective control of approximation levels demonstrated on complex benchmarks

Machine learning models accurately predict cost and error metrics

Optimized knob settings improve resource efficiency while maintaining desired accuracy

Abstract

Approximate computing trades off accuracy of results for resources such as energy or computing time. There is a large and rapidly growing literature on approximate computing that has focused mostly on showing the benefits of approximation. However, we know relatively little about how to control approximation in a disciplined way. This document briefly describes our published work of controlling approximation for non-streaming programs that have a set of "knobs" that can be dialed up or down to control the level of approximation of different components in the program. The proposed system, Capri, solves this control problem as a constrained optimization problem. Capri uses machine learning to learn cost and error models for the program, and uses these models to determine, for a desired level of approximation, knob settings that optimize metrics such as running time or energy usage. Experimental results with complex benchmarks from different problem domains demonstrate the effectiveness of this approach. This report outlines improvements and extensions to the existing Capri system to address its limitations, including a complete rewrite of the software, and discusses directions for follow up work. The document also includes instructions and guidelines for using the new Capri infrastructure.