On application of heterogeneous hardware architectures based on advanced risc machines to FEM calculations

by

On application of heterogeneous hardware architectures based on advanced risc machines to FEM calculations

Krzysztof Bzowski, Łukasz Rauch

AGH – University of Science and Technology, Mickiewicza 30, 30-059 Kraków, Poland.

DOI:

https://doi.org/10.7494/cmms.2015.3.0548

Abstract:

Advanced RISC Machine (ARM) hardware architectures are nowadays one of the most popular solutions among processors widely present in mobile and embedded systems. Due to relatively low power consumption and high multithreaded capabilities they can be found in more than 75% of 32-bits devices (Frenzel Jr, 2010). Modern ARM processors also contain integrated high efficiency graphics units like Mali T6xx which made them particularly useful for growing market of mobile devices. Mali processors support OpenCL standard which made them valuable for wide range of scientific computing, where processing power is as much important as power consumption. Presented paper contains proof of concept of Finite Element Method (FEM) software capable to compute transient heat transfer analysis and implemented for ARM architecture. Exemplary implementation using OpenCL was prepared. Efficiency data as well as comparison between modern GPGPU, accelerators and ARM devices are included in the paper.

Cite as:

Bzowski, K., & Rauch, Ł. (2015). On application of heterogeneous hardware architectures based on advanced risc machines to FEM calculations. Computer Methods in Materials Science, 15(3), 435-440. https://doi.org/10.7494/cmms.2015.3.0548

Article (PDF):

Keywords:

Heterogeneous computing, GPGPU, ARM, Finite element method

Publication dates:

Received: 11.12.2014, accepted: 13.05.2015, published:

Publication type:

Original scientific paper

References:

Abdurachmanov, D., Bockelman, B., Elmer, P., Eulisse, G., Knight, R., Muzaffar, S., 2014a, Heterogeneous High Throughput Scientific Computing with APM X-Gene and Intel Xeon Phi, ArXiv e-prints, available online at http://arxiv.org/pdf/1410.3441, accessed: 12.10.2015.

Abdurachmanov, D., Elmer, P., Eulisse, G., Muzaffar, S., 2014b, Initial explorations of ARM processors for scientific computing, Journal of Physics Conference Series, 523, 1-6. C N

Bientinesi, P., Herrero, J.R., Quintana-Ortí, E.S., Strzodka, R., C 2015, Parallel computing on graphics processing units and heterogeneous platforms, Concurrency and Computation: Practice and Experience, 27, 1525-1527. A R Frenzel Jr, L.E., 2010, Chapter – How Microcomputers Work: The Brains of Every Electronic Product Today, Elec- A M tronics Explained, ed. Frenzel, L.E., Newnes, Boston, N 123-145.

Grasso, I., Radojkovic, P., Rajovic, N., Gelado, I., Ramirez, A., D O 2014, Energy Efficient HPC on Embedded SoCs: Opti- H mization Techniques for Mali GPU, Proc. Conf. Parallel and Distributed Processing Symposium, 2014 IEEE 28th M International, 123-132. R

Kreutzer, M., Hager, G., Wellein, G., Fehske, H., Bishop, A.R., U 2014, A unified sparse matrix data format for efficient P M general sparse matrix-vector multiply on modern proces- O C sors with wide SIMD units, ArXiv e-prints, available online at http://arxiv.org/pdf/1307.6209v2.pdf, accessed: 12.10.2015. –

Kruze, F., Banas, K., 2014, Finite Element Numerical Integra- tion on Xeon Phi coprocessor, eds. Ganzha, M.,

Maciaszek, L., Paprzycki, M., Proc. Conf. Federated Conference on Computer Science and Information Systems, Lodz, 603-612.

Rajovic, N., Rico, A., Puzovic, N., Adeniyi-Jones, C., Ramirez, A., 2014, Tibidabo: Making the case for an ARM-based HPC system, Future Generation Computer Systems, 36, 322-334.

Rauch, L., 2013, Heterogeneous Hardware Implementation of Molecular Static Method for Modelling of Interatomic Behaviour, Procedia Computer Science, 18, 1057-1067.

Rauch, L., Bzowski, K., Rodzaj, A., 2012, OpenCL Implementation of Cellular Automata Finite Element (CAFE) Method, Parallel Processing and Applied Mathematics, eds. Wyrzykowski, R., Dongarra, J., Karczewski, K.,

Waśniewski, J., Springer Berlin Heidelberg, 381-390.

Rodero, I., Parashar, M., 2012, Energy Efficiency in HPC Systems, Energy-Efficient Distributed Computing Systems, John Wiley & Sons, Inc., 81-108.

Rupp, K., Rudolf, F., Weinbub, J., 2010, ViennaCL – A High Level Linear Algebra Library for GPUs and Multi-Core CPUs, eds. Mehofer, E., Schordan, M., Quinlan, D., Di

Martino, B., Proc. Conf. International Workshop on GPUs and Scientific Applications (GPUScA 2010), 51- 56.