Home // International Journal On Advances in Systems and Measurements, volume 2, number 4, 2009 // View article


System Level Analysis for Achieving Thermal Balance and Lifetime Reliability in Reliably Overclocked Systems

Authors:
Prem Kumar Ramesh
Viswanathan Subramanian
Arun K. Somani

Keywords: Microprocessors, Dependability, Adaptability, Overclocking, Thermal Throttling

Abstract:
Advancements in process technology offer continuous improvements in system performance. Technology scaling brings forth several new challenges. In particular, process, voltage, and temperature variations require sufficient safety margins to be added to the clock frequency of digital systems, making it overly conservative. Aggressive, but reliable, dynamic clock frequency tuning mechanisms that achieve higher system performance, by adapting the clock rates beyond worst-case limits, have been proposed earlier. Even though reliable overclocking guarantees functional correctness, it leads to higher power consumption and overheating. As a consequence, reliable overclocking without considering on-chip temperatures will bring down the lifetime reliability of the chip. In [1], we presented a comparative study on the thermal behavior of reliably overclocked systems with non-accelerated systems. In this paper, we elaborate more on the theoretical analysis along with experimental results to establish a safe acceleration zone for such ‘better than worst-case’ designs by efficiently balancing the gains of overclocking and the impact on system temperature. We analyze how reliable overclocking impacts the on-chip temperature of microprocessors, and evaluate the effects of overheating, due to reliable dynamic overclocking mechanisms, on the lifetime reliability of such systems. First, we theoretically study the possibilities for realizing such a system. We, then, evaluate the effects of thermal throttling, a technique that clamps the on-chip temperature below a predefined value, on system performance and reliability. Our study shows that a reliably overclocked system with dynamic thermal throttling, constrained to operating within 355K, achieves around 25% performance improvement.

Pages: 258 to 268

Copyright: Copyright (c) to authors, 2009. Used with permission.

Publication date: March 17, 2010

Published in: journal

ISSN: 1942-261x