At SMACD 2018

  • Name:

    Design for Reliability: From Devices to Systems

  • Venue:

    15th International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design

  • Date:

    July 2-5, 2018

  • Project Leader:

    Hussam Amrouch, et al.

  • Speaker:

    Hussam Amrouch, Montserrat Nafría, Elisenda Roca

  • Throughout technology scaling, sustaining reliability has become extraordinary challenging. Key to the design of reliable circuits and systems is to account for the uncertainty introduced in the electrical characteristics of devices by Time Zero Variability (TZV), also known as process variations, and Time Dependent Variability (TDV) phenomena (such as Random Telegraph Noise, Bias Temperature Instabilities and Channel Hot Carriers Injection) and its impact on circuits and systems. This tutorial will present the newest research results in reliability from ground up; starting from physics where reliability degradations do originate all the way up to the system level where their ultimate impact finally takes place.

    In the first part of the tutorial, the physics behind TDV phenomena, the characterization challenges in advanced nodes and their compact modeling will be addressed. The second part of the tutorial will use these stochastic compact models to build circuit simulators accounting for this reliability information. They constitute the cornerstone of new reliability-aware design methodologies and tools that help designers to optimize the circuit lifetime. In the last part of the tutorial, we will demonstrate how reliability awareness can be brought to the existing EDA tool flows, which enables designers to proactively build more resilient circuits. This is done through so-called reliability-aware cell libraries. We will show how using these libraries will allow designers to also quantify the ultimate impact of aging at the system level. Finally, we will explain how approximate computing principles can be explored in the context of aging towards maximizing the efficiency.