SlackProbe: A Low Overhead In Situ On-line Timing Slack Monitoring Methodology

Abstract

Abstract-In situ monitoring is an accurate way to monitor circuit delay or timing slack, but usually incurs significant overhead. We observe that most existing slack monitoring methods exclusively focus on monitoring path ending registers, which is not cost efficient from power and area perspectives.In this paper, we propose SlackProbe methodology, which inserts timing slack monitors like "probes" at a selected set of nets, including intermediate nets along critical paths. SlackProbe can significantly reduce the total number of monitors required at the cost of some additional delay margin. It can be used to detect impending delay failures due to various reasons (process variations, ambient fluctuations, circuit aging, etc.) and can be used with various preventive actions (e.g. voltage/frequency scaling, clock stretching/time borrowing, etc.). Though we focus on monitor selection in this work, we give an example of using SlackProbe with adaptive voltage scaling.Experimental results on commercial processors show that with 5% more timing margin, SlackProbe can reduce the number of monitors by 15-18X as compared to the number of monitors inserted at path ending pins.

A cross-layer SER analysis in the presence of PVTA variations

“…Process, Voltage, Temperature, and Aging (PVTA) variation is one of the major dark sides of the technology scaling [19][20][21][22][23][24]. Due to the process variations [25,26] attributes of transistors (such as gate length, oxide thickness, diffusion depth, and V th variation mainly because of random dopant fluctuation) deviate from their nominal value.…”

Accurate and inexpensive performance monitoring for variability-aware systems

“…Design-dependent ring oscillator (DDRO) [15] designs and leverages multiple replicas to accurately monitor circuit performance. SlackProbe [16] inserts in situ monitors at both path intermediate nets and path endpoints, which can greatly reduce the overhead of in situ monitoring. Then we present our implementation of an end-to-end variability-aware system [17] which utilizes the hardware monitors and operates with both hardware and software adaptation.…”

SlackProbe: A Flexible and Efficient In Situ Timing Slack Monitoring Methodology

“…Another possible path selection method is based on the process corners as described in [18], where paths are selected based on their timing slacks at particular process corners. Other path selection methods such as statistical methods [23] can also be used in SlackProbe depending on the specific applications.…”

“…In [18], a cost function c i for each net n i is defined, and combines the monitor power cost and design perturbation cost as:…”

“…SlackProbe monitors in situ timing slack of selected circuit nets, including intermediate nets along circuit paths, which is more power and area efficient. This paper is an extension of [18]. The key contributions of our paper include the following: 1) We propose a novel slack monitoring methodology allowing placing monitors at intermediate nets along circuit paths 2) We develop a metric named "opportunism window", which allows us to flexibly select the set of critical paths to be monitored.…”