Approximate Computing: From Circuits to Applications

Volume 108, Issue 12 | December 2020

Guest Editors:

Weiqiang Liu
Fabrizio Lombardi
Michael Schulte

Special Issue Papers

Scanning the Issue

By W. Liu, F. Lombardi, and M. Schulten

Approximate Arithmetic Circuits: A Survey, Characterization, and Recent Applications

By H. Jiang, F. J. H. Santiago, H. Mo, L. Liu, and J. Han

This article provides a comprehensive evaluation of recently proposed approximate arithmetic circuits mainly including adders, multipliers, and dividers; they are compared under different design constraints and applied to image processing and deep learning applications.

Elementary Functions and Approximate Computing

By J.-M. Muller

This article presents classical approaches of approximate elementary functions with the mainstream techniques of shift-and-add algorithms, polynomial or rational approximations, table-based methods, and bit manipulation.

Circuit-Level Techniques for Logic and Memory Blocks in Approximate Computing Systemsx

By S. Amanollahi, M. Kamal, A. Afzali-Kusha, and M. Pedram

This article reviews the circuit-level techniques for both approximate logic and memory blocks.

A Survey of Testing Techniques for Approximate Integrated Circuits

By M. Traiola, A. Virazel, P. Girard, M. Barbareschi, and A. Bosio

This article examines the test procedure of approximate integrated circuits by identifying the main approximation-aware testing phases.

Approximate Logic Synthesis: A Survey

By I. Scarabottolo, G. Ansaloni, G. A. Constantinides, L. Pozzi, and S. Reda

This article reviews the transformation methods for functional approximation, which can achieve an approximate Boolean function from its exact designs.

Security in Approximate Computing and Approximate Computing for Security: Challenges and Opportunities

By W. Liu, C. Gu, M. O’Neill, G. Qu, P. Montuschi, and F. Lombardi

This article focuses on an emerging area that links security and approximate computing.

Efficient AI System Design With Cross-Layer Approximate Computing

By S. Venkataramani et al.

This article presents IBM’s RAPID, which is a multi-TOPs DNN hardware accelerator core fabricated using 14-nm technology.

Deep In-Memory Architectures in SRAM: An Analog Approach to Approximate Computing

By M. Kang, S. K. Gonugondla, and N. R. Shanbhag

This article provides an overview of the recently proposed deep in-memory architectures (DIMAs) with several approximate prototype chips using 65-nm technology for hardware acceleration of machine learning algorithms.

Resistive Crossbars as Approximate Hardware Building Blocks for Machine Learning: Opportunities and Challenges

By I. Chakraborty, M. Ali, A. Ankit, S. Jain, S. Roy, S. Sridharan, A. Agrawal, A. Raghunathan, and K. Roy

This article presents a comprehensive overview of the emerging paradigm of approximate computing using NVM crossbars for accelerating machine learning workloads.