Dataset for Enhancing State Retention With Energy-Efficient Memory Tracing in Intermittent Systems
Dataset for Enhancing State Retention With Energy-Efficient Memory Tracing in Intermittent Systems
This dataset supports the publication "Enhancing State Retention with Energy-Efficient Memory Tracing in Intermittent Systems," to be published in the IEEE Transactions on Sustainable Computing (T-SUSC).
DOI: 10.1109/TSUSC.2025.3621509
The dataset contains experimental results from the proposed memory tracing system (MeTra) for energy-efficient state retention in intermittently powered embedded platforms. Results include behavior during different workloads, granularity settings, RAM-to-NVM transfer times, energy measurements, and FPGA resource utilization with and without MeTra enabled.
All data are provided in a single Excel file (MeTra_Dataset.xlsx). Each sheet corresponds to a figure or table presented in the paper:
Sheet ‘Fig. 5’ — Data supporting Fig. 5: Behaviour of MeTra when running matrix multiplication with a RAM block size (granularity value) of 4 bytes, where 18.21% of the RAM was utilized, showing voltage threshold updates after each counter overflow interrupt.
Sheet ‘Fig. 6’ — Data supporting Fig. 6: Main memory (RAM) utilization and corresponding save time to NVM for granularity values of 4 and 8 bytes.
Sheet ‘Fig. 7’ — Data supporting Fig. 7: Comparative analysis of trace memory usage and RAM-to-NVM save duration at various stages of computational progress (% of application run) for matrix multiplication, array sorting, and AES, evaluated at granularities of 4 and 8 bytes.
Sheet ‘Fig. 8’ — Data supporting Fig. 8: Comparative analysis of trace memory usage and RAM-to-NVM save duration across multiple iterations of matrix multiplication, array sorting, and AES at granularities of 4 and 8 bytes.
Sheet ‘Fig. 9’ — Data supporting Fig. 9: Energy consumption to save selected RAM to NVM during state saving.
Sheet ‘Table II’ — Data supporting TABLE II: Resource utilization by the system, with and without the memory tracing system (MeTra).
File format: Microsoft Excel Workbook (.xlsx)
Licence: CC BY 4.0
University of Southampton
Bin Tariq, Osama
67a8d20d-9aa2-4bd5-b0c0-3d8cb019ca20
Verykios, Theodoros
fc203333-af9c-48e6-b7d6-f22d8cf60636
Merrett, Geoff
89b3a696-41de-44c3-89aa-b0aa29f54020
Balsamo, Domenico
9cfdb7ce-3fa9-49a5-b119-77897d6db64d
Bin Tariq, Osama
67a8d20d-9aa2-4bd5-b0c0-3d8cb019ca20
Verykios, Theodoros
fc203333-af9c-48e6-b7d6-f22d8cf60636
Merrett, Geoff
89b3a696-41de-44c3-89aa-b0aa29f54020
Balsamo, Domenico
9cfdb7ce-3fa9-49a5-b119-77897d6db64d
Bin Tariq, Osama, Verykios, Theodoros, Merrett, Geoff and Balsamo, Domenico
(2025)
Dataset for Enhancing State Retention With Energy-Efficient Memory Tracing in Intermittent Systems.
University of Southampton
doi:10.5258/SOTON/D3726
[Dataset]
Abstract
This dataset supports the publication "Enhancing State Retention with Energy-Efficient Memory Tracing in Intermittent Systems," to be published in the IEEE Transactions on Sustainable Computing (T-SUSC).
DOI: 10.1109/TSUSC.2025.3621509
The dataset contains experimental results from the proposed memory tracing system (MeTra) for energy-efficient state retention in intermittently powered embedded platforms. Results include behavior during different workloads, granularity settings, RAM-to-NVM transfer times, energy measurements, and FPGA resource utilization with and without MeTra enabled.
All data are provided in a single Excel file (MeTra_Dataset.xlsx). Each sheet corresponds to a figure or table presented in the paper:
Sheet ‘Fig. 5’ — Data supporting Fig. 5: Behaviour of MeTra when running matrix multiplication with a RAM block size (granularity value) of 4 bytes, where 18.21% of the RAM was utilized, showing voltage threshold updates after each counter overflow interrupt.
Sheet ‘Fig. 6’ — Data supporting Fig. 6: Main memory (RAM) utilization and corresponding save time to NVM for granularity values of 4 and 8 bytes.
Sheet ‘Fig. 7’ — Data supporting Fig. 7: Comparative analysis of trace memory usage and RAM-to-NVM save duration at various stages of computational progress (% of application run) for matrix multiplication, array sorting, and AES, evaluated at granularities of 4 and 8 bytes.
Sheet ‘Fig. 8’ — Data supporting Fig. 8: Comparative analysis of trace memory usage and RAM-to-NVM save duration across multiple iterations of matrix multiplication, array sorting, and AES at granularities of 4 and 8 bytes.
Sheet ‘Fig. 9’ — Data supporting Fig. 9: Energy consumption to save selected RAM to NVM during state saving.
Sheet ‘Table II’ — Data supporting TABLE II: Resource utilization by the system, with and without the memory tracing system (MeTra).
File format: Microsoft Excel Workbook (.xlsx)
Licence: CC BY 4.0
Spreadsheet
MeTra_Dataset.xlsx
- Dataset
Restricted to Registered users only
Text
MeTra_Dataset_Readme.txt
- Text
More information
Published date: 2025
Identifiers
Local EPrints ID: 506155
URI: http://eprints.soton.ac.uk/id/eprint/506155
PURE UUID: 82219565-ebf3-4ce5-a428-aee0aefd5a0a
Catalogue record
Date deposited: 29 Oct 2025 17:39
Last modified: 30 Oct 2025 02:38
Export record
Altmetrics
Contributors
Creator:
Osama Bin Tariq
Creator:
Theodoros Verykios
Creator:
Geoff Merrett
Creator:
Domenico Balsamo
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