Modern multiprocessor systems-on-chip (MpSoCs) offer tremendous power and performance optimization opportunities by tuning thousands of potential voltage, frequency and core configurations. As the workload phases change at runtime, different configurations may become optimal with respect to power, performance or other metrics. Identifying the optimal configuration at runtime is infeasible due to the large number of workloads and configurations. This paper proposes a novel methodology that can find the Pareto-optimal configurations at runtime as a function of the workload. To achieve this, we perform an extensive offline characterization to find classifiers that map performance counters to optimal configurations. Then, we use these classifiers and performance counters at runtime to choose Pareto-optimal configurations. We evaluate the proposed methodology by maximizing the performance per watt for 18 single- and multi-threaded applications. Our experiments demonstrate an average increase of 93%, 81% and 6% in performance per watt compared to the interactive, ondemand and powersave governors, respectively.
DyPO: Dynamic Pareto-Optimal Configuration Selection for Heterogeneous MpSoCs
- | Published On:
Chetan Arvind Patil
Hi, I am Chetan Arvind Patil (chay-tun – how to pronounce), a semiconductor professional whose job is turning data into products for the semiconductor industry that powers billions of devices around the world. And while I like what I do, I also enjoy biking, working on few ideas, apart from writing, and talking about interesting developments in hardware, software, semiconductor and technology.
COPYRIGHT 2024, CHETAN ARVIND PATIL
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. In other words, share generously but provide attribution.
DISCLAIMER
Opinions expressed here are my own and may not reflect those of others. Unless I am quoting someone, they are just my own views.
RECENT POSTS
DALL-E Continuous Package-Level Scaling Moore’s Law, which predicted the doubling of transistors every two years, is slowing down due
DALL-E Semiconductor Defect Pattern Analysis The relentless drive for higher performance, lower power consumption, and increased functionality in semiconductor
DALL-E The Growing Need For High-Speed Interconnects As AI workloads grow exponentially, the demand for faster, more efficient interconnects
DALL-E Specialized Process The semiconductor industry is no longer defined solely by the race for smaller nodes, such as
DALL-E What Is Hybrid AI? Hybrid AI represents a paradigm shift in AI architecture. Instead of relying solely on
