By Topic

Predicting thermal behavior for temperature management in time-critical multicore systems

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

3 Author(s)
Buyoung Yun ; EECS Dept., Univ. of Michigan, Ann Arbor, MI, USA ; Shin, K.G. ; Shige Wang

Multi-core System-on-Chip (SoC) has become a popular execution platform for many embedded real-time systems that require high performance and low power-consumption. High temperature is known to accelerate the failure of deep submicron chips. To prevent such accelerated failures due to chip overheating, various thermal-aware scheduling (TAS) algorithms and dynamic thermal management (DTM) have been proposed and applied to mission/safety-critical applications. To control on-chip temperature more effectively, it is necessary to predict the thermal dynamics of a multi-core chip in real time and trigger appropriate power/temperature management before overheating the chip. However, due to dynamically-changing runtime environments, it is very difficult to estimate the chip temperature on-the-fly. In this paper, we propose models of efficiently estimating multi-core chip temperature while accounting for the system dynamics in real time. Based on these models, we design a proactive peak temperature manager (PTM) which periodically estimates future core temperature and triggers proper DTMs on the estimated-to-be-overheated cores for their cooling without violating applications timing constraints. Our in-depth evaluation based on the HotSpot thermal simulator has shown that the proposed method can predict the occurrence of peak temperature in a core with 90-98% accuracy, and using the estimated thermal model of a multi-core chip, PTM can effectively keep core temperature below a given threshold without violating any timing constraint.

Published in:

Real-Time and Embedded Technology and Applications Symposium (RTAS), 2013 IEEE 19th

Date of Conference:

9-11 April 2013