Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation
Schuster, N. 2025. Energy-Efficient Processing in Radio Astronomy: Comparative Analysis of Computing Platforms and their Implementation. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/bf939b00-9e5b-4fd2-ad7e-22c73f381666
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867614035102400512 |
|---|---|
| access_status_str | Open Access |
| author | Schuster, Nikola |
| author2 | Smit, W. A. |
| author_browse | Schuster, Nikola Smit, W. A. |
| author_facet | Smit, W. A. Schuster, Nikola |
| author_sort | Schuster, Nikola |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Schuster, N. 2025. Energy-Efficient Processing in Radio Astronomy: Comparative Analysis of Computing Platforms and their Implementation. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/bf939b00-9e5b-4fd2-ad7e-22c73f381666 |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/132439 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:45:37.487Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/132439 Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation Schuster, Nikola Smit, W. A. Grobler, T. L. Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering. Radio astronomy -- Data processing Computing platforms Field programmable gate arrays Signal processing UCTD Schuster, N. 2025. Energy-Efficient Processing in Radio Astronomy: Comparative Analysis of Computing Platforms and their Implementation. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/bf939b00-9e5b-4fd2-ad7e-22c73f381666 Thesis (MEng)--Stellenbosch University, 2025. ENGLISH ABSTRACT: This thesis investigates the energy efficiency of some high-performance computing hardware used for radio astronomy, with a focus on Field-Programmable Gate Arrays (FPGAs) and Graphics Processing Units (GPUs), and how the FPGAs are programmed. The Square Kilometer Array (SKA), a groundbreaking radio astronomy project, requires advanced computational hardware to handle vast volumes of data, making energy efficiency a critical concern. With limited research backing up the current move to replace some FPGAs, used for processing data on site by the SKA, with GPUs, this research addresses two main objectives: a comparative analysis of energy consumption between FPGAs and GPUs, and investigating the impact of using an abstraction layer (HDL Coder) to write VHDL code, on FPGA energy efficiency. In the first part of the investigation, we ran a correlator algorithm on both FPGA and GPU platforms, measuring power consumption. Power usage was assessed through current and voltage readings. This was used to calculate the instructions per kilowatt-hour which was used to compare the platforms. From this investigation we saw that the FPGA processed 127 times more instructions than the GPU. The second part involved creating a Vivado project from a MATLAB file using the HDL Coder application - an abstraction layer used by the SKA. This VHDL code, along with a self-written script for a correlator was then implemented into a Vivado project, with separate custom IP blocks communicating via the AXI4 bus. The power consumption was measured when the FPGA was running the correlator. After analyzing the data, we found that the project with the VHDL script written by HDL coder consumed 11.18% more energy. Despite advancements in hardware and a shift from FPGAs to GPUs by the SKA, the energy implications of this transition have not been researched thoroughly. Additionally, while abstraction layers like HDL Coder simplify hardware programming, their effect on energy consumption in the context of radio astronomy is not well researched. This thesis aims to bridge these knowledge gaps, offering insights into optimal computing solutions for energy-efficient data processing in radio astronomy. AFRIKAANSE OPSOMMING: Hierdie tesis ondersoek die energiedoeltreffendheid van sekere ho¨eprestasierekenaartoerusting wat in radioastronomie gebruik word, met ’n fokus op FieldProgrammable Gate Arrays (FPGAs) en Graphics Processing Units (GPUs), en hoe die FPGAs geprogrammeer word. Die Square Kilometer Array (SKA), ’n baanbrekende radioastronomieprojek, vereis gevorderde rekenaartoerusting om groot volumes data te hanteer, wat energiedoeltreffendheid ’n kritieke bekommernis maak. Met beperkte navorsing wat die huidige skuif deur die SKA om sekere FPGAs met GPUs te vervang ondersteun, spreek hierdie navorsing twee hoofdoelwitte aan: ’n vergelykende analise van energieverbruik tussen FPGAs en GPUs, en die ondersoek van die impak van die gebruik van ’n abstraksielaag (HDL Coder) om VHDL-kode te skryf, op FPGA-energiedoeltreffendheid. In die eerste deel van die ondersoek het ons ’n korrelator-algoritme op beide FPGA- en GPU-platforms laat loop, terwyl ons kragverbruik gemeet het. Kragverbruik is beoordeel deur stroom- en spanningslesings, wat gebruik is om die instruksies per kilowatt-uur te bereken en die platforms te vergelyk. Uit hierdie ondersoek het ons gesien dat die FPGA 127 keer meer instruksies verwerk het as die GPU. Die tweede deel het behels om ’n Vivado-projek vanaf ’n MATLAB-lˆeer te skep met behulp van die HDL Coder-toepassing - ’n abstraksielaag wat deur die SKA gebruik word. Hierdie VHDL-kode, saam met ’n self-geskrewe skrip vir ’n korrelator, is toe in aparte pasgemaakte IP-blokke ge¨ımplementeer wat via die AXI4-bus kommunikeer. Die kragverbruik is gemeet terwyl die FPGA die korrelator bedryf het. Ons het gevind dat die projek met die VHDL-skrif geskryf deur HDL Coder 11.18 % meer krag verbruik het. Ten spyte van vooruitgang in hardeware en ’n skuif van FPGAs na GPUs deur die SKA, is die energie-implikasies van hierdie oorgang nog nie deeglik nagevors nie. Daarbenewens, hoewel abstraksielae soos HDL Coder hardewareprogrammering vereenvoudig, is hul effek op energieverbruik in die konteks van radioastronomie nie goed nagevors nie. Hierdie tesis het ten doel om hierdie kennisgapings te oorbrug en insigte te bied in optimale rekenaartoepassings vir energiedoeltreffende dataverwerking in radioastronomie. Masters 2025-06-06T13:43:23Z 2025-06-06T13:43:23Z 2025-03 Thesis https://scholar.sun.ac.za/handle/10019.1/132439 en Stellenbosch University xvi, 127 pages : illustrations application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Radio astronomy -- Data processing Computing platforms Field programmable gate arrays Signal processing UCTD Schuster, Nikola Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation |
| title | Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation |
| title_full | Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation |
| title_fullStr | Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation |
| title_full_unstemmed | Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation |
| title_short | Energy-efficient processing in radio astronomy: comparative analysis of computing platforms and their implementation |
| title_sort | energy efficient processing in radio astronomy comparative analysis of computing platforms and their implementation |
| topic | Radio astronomy -- Data processing Computing platforms Field programmable gate arrays Signal processing UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/132439 |
| work_keys_str_mv | AT schusternikola energyefficientprocessinginradioastronomycomparativeanalysisofcomputingplatformsandtheirimplementation |