A4 Refereed article in a conference publication
Execution Frequency and Energy Optimization for DVFS-enabled, Near-threshold Processors
Authors: Mäkikyrö Sofia, Tuoriniemi Samuli, Anttila Risto, Koskinen Lauri
Editors: IEEE
Conference name: International Conference on Advanced Computer Information Technologies (ACIT)
Publication year: 2020
Book title : 2020 10th International Conference on Advanced Computer Information Technologies (ACIT'2020)
Journal name in source: 2020 10TH INTERNATIONAL CONFERENCE ON ADVANCED COMPUTER INFORMATION TECHNOLOGIES (ACIT)
First page : 518
Last page: 522
Number of pages: 5
ISBN: 978-1-7281-6759-6
eISBN: 978-1-7281-6760-2
DOI: https://doi.org/10.1109/ACIT49673.2020.9208896
Abstract
Shown here is an analysis algorithm targeted for near-threshold ultra-wide Dynamic Voltage and Frequency Scaling (UW-DVFS) embedded systems. The algorithm defines execution frequencies based on the software execution trace. Task execution profiling and constraints are used to form a linear optimization problem. Execution frequencies are defined based on the optimization result, the DVFS scaling factors. The system's DVFS overhead is measured and included in the model. The analyzed code is tested on a near-threshold ARM Cortex-M3 core with integrated PLL and power management. In this case, over 49% energy savings can be achieved for industry-standard speech-recognition software without any penalty in application throughput.
Shown here is an analysis algorithm targeted for near-threshold ultra-wide Dynamic Voltage and Frequency Scaling (UW-DVFS) embedded systems. The algorithm defines execution frequencies based on the software execution trace. Task execution profiling and constraints are used to form a linear optimization problem. Execution frequencies are defined based on the optimization result, the DVFS scaling factors. The system's DVFS overhead is measured and included in the model. The analyzed code is tested on a near-threshold ARM Cortex-M3 core with integrated PLL and power management. In this case, over 49% energy savings can be achieved for industry-standard speech-recognition software without any penalty in application throughput.
UTU Research Portal