Hardware acceleration of reaction-diffusion systems: a guide to optimisation of pattern formation algorithms using OpenACC

Ruth E. Falconer, Alasdair N. Houston, Xavier Portell, Wilfred Otten

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    39 Downloads (Pure)

    Abstract

    Reaction Diffusion Systems (RDS) have widespread applications in computational ecology, biology, computer graphics and the visual arts. For the former applications a major barrier to the development of effective simulation models is their computational complexity - it takes a great deal of processing power to simulate enough replicates such that reliable conclusions can be drawn. Optimizing the computation is thus highly desirable in order to obtain more results with less resources. Existing optimizations of RDS tend to be low-level and GPGPU based. Here we apply the higher-level OpenACC framework to two case studies: a simple RDS to learn the ‘workings’ of OpenACC and a more realistic and complex example. Our results show that simple parallelization directives and minimal data transfer can produce a useful performance improvement. The relative simplicity of porting OpenACC code between heterogeneous hardware is a key benefit to the scientific computing community in terms of speed-up and portability.
    Original languageEnglish
    Title of host publicationSpring Simulation Conference (SpringSim), 2019
    PublisherIEEE
    Pages472-483
    Number of pages12
    ISBN (Electronic)9781510883888
    ISBN (Print)9781728135472
    DOIs
    Publication statusPublished - 10 Jun 2019
    Event2019 Spring Simulation Conference - University of Arizona, Tucson, United States
    Duration: 29 Apr 20192 May 2019
    http://scs.org/springsim/

    Conference

    Conference2019 Spring Simulation Conference
    Abbreviated titleSpringSim'19
    CountryUnited States
    CityTucson
    Period29/04/192/05/19
    Internet address

    Fingerprint

    Hardware
    Natural sciences computing
    Ecology
    Computer graphics
    Data transfer
    Computer hardware
    Computational complexity
    Processing

    Cite this

    Falconer, R. E., Houston, A. N., Portell, X., & Otten, W. (2019). Hardware acceleration of reaction-diffusion systems: a guide to optimisation of pattern formation algorithms using OpenACC. In Spring Simulation Conference (SpringSim), 2019 (pp. 472-483). IEEE . https://doi.org/10.23919/SpringSim.2019.8732883
    Falconer, Ruth E. ; Houston, Alasdair N. ; Portell, Xavier ; Otten, Wilfred. / Hardware acceleration of reaction-diffusion systems : a guide to optimisation of pattern formation algorithms using OpenACC. Spring Simulation Conference (SpringSim), 2019. IEEE , 2019. pp. 472-483
    @inproceedings{e402020c12f54dcb950c9c3d6f57d542,
    title = "Hardware acceleration of reaction-diffusion systems: a guide to optimisation of pattern formation algorithms using OpenACC",
    abstract = "Reaction Diffusion Systems (RDS) have widespread applications in computational ecology, biology, computer graphics and the visual arts. For the former applications a major barrier to the development of effective simulation models is their computational complexity - it takes a great deal of processing power to simulate enough replicates such that reliable conclusions can be drawn. Optimizing the computation is thus highly desirable in order to obtain more results with less resources. Existing optimizations of RDS tend to be low-level and GPGPU based. Here we apply the higher-level OpenACC framework to two case studies: a simple RDS to learn the ‘workings’ of OpenACC and a more realistic and complex example. Our results show that simple parallelization directives and minimal data transfer can produce a useful performance improvement. The relative simplicity of porting OpenACC code between heterogeneous hardware is a key benefit to the scientific computing community in terms of speed-up and portability.",
    author = "Falconer, {Ruth E.} and Houston, {Alasdair N.} and Xavier Portell and Wilfred Otten",
    year = "2019",
    month = "6",
    day = "10",
    doi = "10.23919/SpringSim.2019.8732883",
    language = "English",
    isbn = "9781728135472",
    pages = "472--483",
    booktitle = "Spring Simulation Conference (SpringSim), 2019",
    publisher = "IEEE",

    }

    Falconer, RE, Houston, AN, Portell, X & Otten, W 2019, Hardware acceleration of reaction-diffusion systems: a guide to optimisation of pattern formation algorithms using OpenACC. in Spring Simulation Conference (SpringSim), 2019. IEEE , pp. 472-483, 2019 Spring Simulation Conference, Tucson, United States, 29/04/19. https://doi.org/10.23919/SpringSim.2019.8732883

    Hardware acceleration of reaction-diffusion systems : a guide to optimisation of pattern formation algorithms using OpenACC. / Falconer, Ruth E.; Houston, Alasdair N.; Portell, Xavier; Otten, Wilfred.

    Spring Simulation Conference (SpringSim), 2019. IEEE , 2019. p. 472-483.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    TY - GEN

    T1 - Hardware acceleration of reaction-diffusion systems

    T2 - a guide to optimisation of pattern formation algorithms using OpenACC

    AU - Falconer, Ruth E.

    AU - Houston, Alasdair N.

    AU - Portell, Xavier

    AU - Otten, Wilfred

    PY - 2019/6/10

    Y1 - 2019/6/10

    N2 - Reaction Diffusion Systems (RDS) have widespread applications in computational ecology, biology, computer graphics and the visual arts. For the former applications a major barrier to the development of effective simulation models is their computational complexity - it takes a great deal of processing power to simulate enough replicates such that reliable conclusions can be drawn. Optimizing the computation is thus highly desirable in order to obtain more results with less resources. Existing optimizations of RDS tend to be low-level and GPGPU based. Here we apply the higher-level OpenACC framework to two case studies: a simple RDS to learn the ‘workings’ of OpenACC and a more realistic and complex example. Our results show that simple parallelization directives and minimal data transfer can produce a useful performance improvement. The relative simplicity of porting OpenACC code between heterogeneous hardware is a key benefit to the scientific computing community in terms of speed-up and portability.

    AB - Reaction Diffusion Systems (RDS) have widespread applications in computational ecology, biology, computer graphics and the visual arts. For the former applications a major barrier to the development of effective simulation models is their computational complexity - it takes a great deal of processing power to simulate enough replicates such that reliable conclusions can be drawn. Optimizing the computation is thus highly desirable in order to obtain more results with less resources. Existing optimizations of RDS tend to be low-level and GPGPU based. Here we apply the higher-level OpenACC framework to two case studies: a simple RDS to learn the ‘workings’ of OpenACC and a more realistic and complex example. Our results show that simple parallelization directives and minimal data transfer can produce a useful performance improvement. The relative simplicity of porting OpenACC code between heterogeneous hardware is a key benefit to the scientific computing community in terms of speed-up and portability.

    U2 - 10.23919/SpringSim.2019.8732883

    DO - 10.23919/SpringSim.2019.8732883

    M3 - Conference contribution

    SN - 9781728135472

    SP - 472

    EP - 483

    BT - Spring Simulation Conference (SpringSim), 2019

    PB - IEEE

    ER -