Performance measurements and comparison between ClústerCruz I and Clúster Cruz II
Keywords:
cluster, raspberry pi versio 2 , parallel computing , distributed computing , MPI, education cluster , portable , comparationAbstract
This paper discusses performance measures (Benchmarking) made in a cluster and comparisons of run time ittakes the cluster CruzI to run certain programs (according to [1] CRUZ means Cluster Research Uni of Zero Power)containing 16 slave nodes and an additional node to the master node with the time it takes the Cluster CruzII 6slave nodes and one master node. The same PI calculation program executed in parallel using MPI functions onboth machines and the time it takes this calculation was measured. The software HPL (High Performance Linpack)was also executed for computing speed CruzII cluster units gigaflops.
Downloads
References
2. Simon J. Cox. Steps to make a Raspberry Pi Supercom- puter. http://www.southampton.ac.uk/sjc/raspberrypi/, 2014.
3. Tina Chen. Instalación del NFS server. http://experimentandoconraspberrypi.blogspot.com/ 2013/06/compartir-directorio-con-nfs.html, 2013.
4. Andrew K. Dennis. Raspberry Pi Super Cluster. 2013.
5. Nikilesh Balakrishnan. Building and benchmarking a low power ARM cluster. 2012.
6. Jimbo Wales. Benchmark. https://es.wikipedia.org/ wiki/Benchmark (inform%C3%A1tica)
7. J. Dongarra A. Cleary A. Petitet, R. C. Whaley. HPL - A Portable Implementation of the High Perfor- mance Linpack Benchmark for Distributed Memory Com- puters. https://www.netlib.org/benchmark/hpl/
8. Akshay Pai. HPL (High Performance Linpack): Benchmarking Raspberry Pis. https://www.howtoforge.com/tutorial/hpl-high- performance-linpack-benchmark-raspberry-pi/
9. Advanced Clustering Technologies. How do I tune my HPL.dat le?. http://www.advancedclustering.com/act-kb/tune- hpl-dat-file/
10. Brian Dye Distributed Computing with the Raspberry Pi, 2014.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2021 REVCIUNI
This work is licensed under a Creative Commons Attribution 4.0 International License.
Articles published by REVCIUNI can be shared through the Creative Commons international public license: CC BY 4.0. Permissions beyond this scope can be consulted through the email revistas@uni.edu.pe