AIRCC's International Journal of Computer Science and Information Technology

Open Access Open Access  Restricted Access Subscription Access

Exploiting Raspberry PI Clusters And Campus Lab Computers For Distributed Computing


Affiliations
1 Department of Electrical and Computer Engineering, Purdue University Fort Wayne Fort Wayne, Indiana, United States
 

Distributed computing networks harness the power of existing computing resources and grant access to significant computing power while averting the costs of a supercomputer. This work aims to configure distributed computing networks using different computer devices and explore the benefits of the computing power of such networks. First, an HTCondor pool consisting of sixteen Raspberry Pi single-board computers and one laptop is created. The second distributed computing network is set up with Windows computers in university campus labs. With the HTCondor setup, researchers inside the university can utilize the lab computers as computing resources. In addition, the HTCondor pool is configured alongside the BOINC installation on both computer clusters, allowing them to contribute to high-throughput scientific computing projects in the research community when the computers would otherwise sit idle. The scalability of these two distributed computing networks is investigated through a matrix multiplication program and the performance of the HTCondor pool is also quantified using its built-in benchmark tool. With such a setup, the limits of the distributed computing network architecture in computationally intensive problems are explored.

Keywords

Distributed Computing, Single-Board Computers, Raspberry Pi.
User
Notifications
Font Size

  • S. S. Vazhkudai, et al. (2018) “The design, deployment, and evaluation of the CORAL pre-exascale systems.” in Proc. Supercomputing 2018 (SC18): 31th Int. Conf. on High Performance Computing, Networking, Storage and Analysis, Dallas, TX.

  • Summit - Oak Ridge leadership computing facility, [online]. Available: https://www.olcf.ornl.gov/summit/ (accessed May 2022).

  • HPE GreenLake for high performance computing platform, [online]. Available: https://www.hpe.com/us/en/greenlake/hpc.html (accessed May 2022).

  • D. P. Anderson, (2020) “BOINC: A platform for volunteer computing,” Journal of Grid Computing, vol. 18, pp. 99-122, doi: 10.1007/s10723-019-09497-9.

  • S. M. Larson, C. D. Snow, M. Shirts, and V. S. Pande, (2009) “Folding@Home and Genome@Home: Using distributed computing to tackle previously intractable problems in computtaional biology,” arXiv preprnt, doi: 10.48550/arXiv.0901.0866.

  • D. Thain, T. Tannenbaum, and M. Livny, (2005) “Distributed computing in practice: The Condor experience,” Concurrency and Computation: Practice and Experience, vol. 17, iss. 2-4, pp. 323-356, doi: 10.1002/cpe.938.

  • H. Mujtaba, “Folding@Home now at almost 2.5 Exaflops to fight COVID-19 – Faster than top 500 supercomputers in the world,” [online]. Available: https://wccftech.com/folding-home-almost-2-5-exaflops-fight-covid-19-faster-than-top-500-world-supercomputers/ (accessed May 2022).

  • A. Petitet, R. C. Whaley, J. Dongarra, A. Cleary, “HPL – A portable implementation of the high-performance linpack benchmark for distributed-memory computers,” [online]. Available: https://www.netlib.org/benchmark/hpl/ (accessed May 2022)

  • M. F. Cloutier, C. Paradis, and V. M. Weaver, (2016) “A Raspberry Pi cluster instrumented for fine-grained power measurement,” Electronics, vol. 5, no. 4, 61, doi: 10.3390/electronics5040061

  • P. J. Basford et al., (2020) “Performance analysis of single board computer clusters,” Future Generation Computer Systems, vol. 102, pp. 278–291, doi: 10.1016/j.future.2019.07.040

  • D. Hawthorne, M. Kapralos, R. W. Blaine, and S. J. Matthews, (2020) “Evaluating cryptographic performance of Rapsberry Pi clusters,” in Proc. 2020 IEEE High Performance Extreme Computing Conference (HPEC), pp. 1-9, doi: 10.1109/HPEC43674.2020.9286247.

  • S. Savazzi, M. Nicoli and V. Rampa, (2020) “Federated Learning with cooperating devices: A consensus approach for massive IoT networks,” IEEE Internet of Things Journal, vol. 7, no. 5, pp. 4641-4654, doi: 10.1109/JIOT.2020.2964162.

  • World Community Grid, [online]. Available: https://www.worldcommunitygrid.org/ (accessed May 2022).

  • HTCondor Overview, [online]. Available: https://htcondor.org/htcondor/overview/ (accessed May 2022).

  • Center for High Throughput Computing, University of Wisconsin–Madison, “Policy Configuration for Execute Hosts and for Submit Hosts — HTCondor Manual 9.4.0 documentation,” [online]. Available: https://htcondor.readthedocs.io/en/latest/admin-manual/policy-configuration.html (accessed December 2021).

  • BOINC: Compute for Science, [online]. Available: https://boinc.berkeley.edu, (accessed May 2022).

  • AMD Ryzen 9 5950X Benchmarks, [online]. Available: https://openbenchmarking.org/vs/Processor/AMD%20Ryzen%209%205950X%2016-Core (accessed December 2021).

  • TOP500 List - November 2021, [online]. Available: https://www.top500.org/lists/top500/list/2021/11/ (accessed May 2022).


Abstract Views: 135

PDF Views: 84




  • Exploiting Raspberry PI Clusters And Campus Lab Computers For Distributed Computing

Abstract Views: 135  |  PDF Views: 84

Authors

Jacob Bushur
Department of Electrical and Computer Engineering, Purdue University Fort Wayne Fort Wayne, Indiana, United States
Chao Chen
Department of Electrical and Computer Engineering, Purdue University Fort Wayne Fort Wayne, Indiana, United States

Abstract


Distributed computing networks harness the power of existing computing resources and grant access to significant computing power while averting the costs of a supercomputer. This work aims to configure distributed computing networks using different computer devices and explore the benefits of the computing power of such networks. First, an HTCondor pool consisting of sixteen Raspberry Pi single-board computers and one laptop is created. The second distributed computing network is set up with Windows computers in university campus labs. With the HTCondor setup, researchers inside the university can utilize the lab computers as computing resources. In addition, the HTCondor pool is configured alongside the BOINC installation on both computer clusters, allowing them to contribute to high-throughput scientific computing projects in the research community when the computers would otherwise sit idle. The scalability of these two distributed computing networks is investigated through a matrix multiplication program and the performance of the HTCondor pool is also quantified using its built-in benchmark tool. With such a setup, the limits of the distributed computing network architecture in computationally intensive problems are explored.

Keywords


Distributed Computing, Single-Board Computers, Raspberry Pi.

References