Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Waterloo

$27,000.00

Scalable and adaptive power management for highly parallel single chip multi-core processors

FY 2012-13

Summary
Highly parallel single chip multi-core processors have become the driving force across product domains ranging from low-power multimedia and mobile devices to chip-multiprocessors used in high-end servers. Such multi-core systems are typically characterized by a high degree of spatio-temporal heterogeneity in application characteristics - from one core to another, and from one application phase to another. In addition, due to the increasing impact of random within-die manufacturing process variations, the leakage power and timing characteristics of otherwise identical cores on the same chip can be vastly different. Finally, due to the increased impact of reliability phenomena like bias temperature instability (BTI) with transistor scaling, the characteristics of the cores on a chip vary with time. In the face of the unpredictability and variability sources mentioned above, guaranteeing (or optimizing for) the power and performance specifications of a multi-core system cannot be done at design-time, and must instead be enabled through fine-grained, run-time adaptive, resource management techniques. The managed resources could be power, performance, temperature, remaining lifetime before failure etc., although in this proposal we focus primarily on power and temperature. The primary contribution of the proposed research is to develop scalable and adaptive control that will enable the next generation of highly parallel multi-core platforms to deliver predictable power/performance in the face of unpredictability and variability arising from the manufacturing process, environment and application behavior. Compared to prior state-of-the-art, the novel contributions of the proposed research are: (i) the focus on the scalability of the adaptive control mechanism, in terms of area, power and performance overheads, to systems with potentially hundreds (or more) cores integrated on the same chip; (ii) holistically addressing all major variation sources within one framework; and (iii) demonstrating the feasibility of the proposed techniques using FPGA-based hardware prototyping.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
402195-2011
Name nom
Garg, Siddharth
Department département
Electrical & Computer Engineering
Organization id
33
Institution établissement
University of Waterloo
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2012
Competition year année de concours
2011
Award amount
27000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
1510
Committee name en
Electrical and Computer Engineering
Committee name fr
Génie électrique et informatique
Area of application code
1207
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Information, computer and communication technologies
Area of application fr
Technologie de l'information, des ordinateurs et des réseaux de communications
Research subject code
2722
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Vlsi systems
Research subject fr
Systèmes intégrés à très grande échelle (VLSI)
Partie
2
Nb partie
5
Application title
Scalable and adaptive power management for highly parallel single chip multi-core processors
Keyword
NSERC
Application summary
Highly parallel single chip multi-core processors have become the driving force across product domains ranging from low-power multimedia and mobile devices to chip-multiprocessors used in high-end servers. Such multi-core systems are typically characterized by a high degree of spatio-temporal heterogeneity in application characteristics - from one core to another, and from one application phase to another. In addition, due to the increasing impact of random within-die manufacturing process variations, the leakage power and timing characteristics of otherwise identical cores on the same chip can be vastly different. Finally, due to the increased impact of reliability phenomena like bias temperature instability (BTI) with transistor scaling, the characteristics of the cores on a chip vary with time. In the face of the unpredictability and variability sources mentioned above, guaranteeing (or optimizing for) the power and performance specifications of a multi-core system cannot be done at design-time, and must instead be enabled through fine-grained, run-time adaptive, resource management techniques. The managed resources could be power, performance, temperature, remaining lifetime before failure etc., although in this proposal we focus primarily on power and temperature. The primary contribution of the proposed research is to develop scalable and adaptive control that will enable the next generation of highly parallel multi-core platforms to deliver predictable power/performance in the face of unpredictability and variability arising from the manufacturing process, environment and application behavior. Compared to prior state-of-the-art, the novel contributions of the proposed research are: (i) the focus on the scalability of the adaptive control mechanism, in terms of area, power and performance overheads, to systems with potentially hundreds (or more) cores integrated on the same chip; (ii) holistically addressing all major variation sources within one framework; and (iii) demonstrating the feasibility of the proposed techniques using FPGA-based hardware prototyping.