Online concurrent workload classification for multi-core energy management
Online concurrent workload classification for multi-core energy management
Modern embedded multi-core processors are organized as clusters of cores, where all cores in each cluster operate at a common Voltage-frequency (V-f ). Such processors often need to execute applications concurrently, exhibiting varying and mixed workloads (e.g. compute- and memory-intensive) depending on the instruction mix and resource sharing. Runtime adaptation is key to achieving energy savings without trading-off application performance with such workload variabilities. In this paper, we propose an online energy management technique that performs concurrent workload classification using the metric Memory Reads Per Instruction (MRPI) and pro-actively selects an appropriate V-f setting through workload prediction. Subsequently, it monitors the workload prediction error and performance loss, quantified by Instructions Per Second (IPS) at runtime and adjusts the chosen V-f to compensate. We validate the proposed technique on an Odroid-XU3 with various combinations of benchmark applications. Results show an improvement in energy efficiency of up to 69% compared to existing approaches.
621-624
Basireddy, Karunakar Reddy
5bfb0b2e-8242-499a-a52b-e813d9a90889
Singh, Amit
bb67d43e-34d9-4b58-9295-8b5458270408
Merrett, Geoff
89b3a696-41de-44c3-89aa-b0aa29f54020
Al-Hashimi, Bashir
0b29c671-a6d2-459c-af68-c4614dce3b5d
23 March 2018
Basireddy, Karunakar Reddy
5bfb0b2e-8242-499a-a52b-e813d9a90889
Singh, Amit
bb67d43e-34d9-4b58-9295-8b5458270408
Merrett, Geoff
89b3a696-41de-44c3-89aa-b0aa29f54020
Al-Hashimi, Bashir
0b29c671-a6d2-459c-af68-c4614dce3b5d
Basireddy, Karunakar Reddy, Singh, Amit, Merrett, Geoff and Al-Hashimi, Bashir
(2018)
Online concurrent workload classification for multi-core energy management.
Design Automation and Test in Europe, , Dresden, Germany.
19 - 23 Mar 2018.
.
Record type:
Conference or Workshop Item
(Paper)
Abstract
Modern embedded multi-core processors are organized as clusters of cores, where all cores in each cluster operate at a common Voltage-frequency (V-f ). Such processors often need to execute applications concurrently, exhibiting varying and mixed workloads (e.g. compute- and memory-intensive) depending on the instruction mix and resource sharing. Runtime adaptation is key to achieving energy savings without trading-off application performance with such workload variabilities. In this paper, we propose an online energy management technique that performs concurrent workload classification using the metric Memory Reads Per Instruction (MRPI) and pro-actively selects an appropriate V-f setting through workload prediction. Subsequently, it monitors the workload prediction error and performance loss, quantified by Instructions Per Second (IPS) at runtime and adjusts the chosen V-f to compensate. We validate the proposed technique on an Odroid-XU3 with various combinations of benchmark applications. Results show an improvement in energy efficiency of up to 69% compared to existing approaches.
Text
Camera_Ready_Version
- Accepted Manuscript
More information
Published date: 23 March 2018
Venue - Dates:
Design Automation and Test in Europe, , Dresden, Germany, 2018-03-19 - 2018-03-23
Identifiers
Local EPrints ID: 415727
URI: http://eprints.soton.ac.uk/id/eprint/415727
PURE UUID: 3b0271f9-9374-4e5d-a1f1-b61660d4f840
Catalogue record
Date deposited: 21 Nov 2017 17:30
Last modified: 16 Mar 2024 03:46
Export record
Contributors
Author:
Karunakar Reddy Basireddy
Author:
Amit Singh
Author:
Geoff Merrett
Author:
Bashir Al-Hashimi
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics