Now showing items 1-13 of 13

  • A checkpoint/restart directive-based approach 

    Maroñas Bravo, Marcos (Universitat Politècnica de Catalunya, 2017)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    Exascale platforms require programming models incorporating support for resilience capabilities since the huge number of components they are expected to have is going to increase the number of errors. Checkpoint/restart ...
  • A Linux Kernel scheduler extension for multi-core systems 

    Roca Nonell, Aleix (Universitat Politécnica de Catalunya, 2017-10-25)
    Master thesis
    Open Access
    In this thesis, it is presented a Linux kernel extension that allows a user-space application to be notified of the blocking and unblocking of its threads, making it possible for a core to execute another thread while the ...
  • Design and development of support for GPU unified memory in OMPSS 

    Rodríguez Soto, Aimar (Universitat Politècnica de Catalunya, 2017-10-31)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    Heterogeneous computing has become prevalent as part of High Performance Computing in the last decade, with asynchronous devices such as Graphics Processing Units constantly evolving. As HPC becomes more specialized and ...
  • Domain Specific Languages for High Performance Computing 

    Fernández Suárez, Alejandro (Universitat Politècnica de Catalunya, 2013-06-21)
    Bachelor thesis
    Open Access
    [ANGLÈS] High Performance Computing (HPC) relies completely on complex parallel, heterogeneus architectures and distributed systems which are hard and error-prone to exploit, even for HPC specialists. Further and further ...
  • Enhanced scheduling techniques through lightweight monitoring for OmpSs-2 

    Navarro Muñoz, Antoni (Universitat Politècnica de Catalunya, 2019-01-28)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    Manually tuning applications to achieve optimal configurations prevails as the default technique. This raises issues, since these configurations depend on many aspects. To solve them, we propose extending runtimes with ...
  • Enhancing the interoperability between distributed-memory and task-based programming models 

    Sala Penadés, Kevin (Universitat Politècnica de Catalunya, 2018-06-27)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    Hybrid applications allow to exploit both inter- and intra-node parallelism, however the programming models currently used are not designed to be combined. For this reason, we propose a generic mechanism to enhance the ...
  • Extending OmpSs-2 with flexible task-based array reductions 

    Pallarès, Ferran (Universitat Politècnica de Catalunya, 2019-01-15)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    Reductions are a well-known computational pattern found in scientific applications that needs efficient parallelisation mechanisms. In this thesis we present a flexible scheme for computing reductions of arrays in the ...
  • Extending Saiph to simulate fluid mechanics and chemistry problems 

    Macià Sorrosal, Sandra (Universitat Politècnica de Catalunya, 2016-02-03)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    Saiph is a domain specific language for solving partial differential equations in high performance computing systems, developed at the Barcelona Computing Center. The aim of this project is to extend this language to support ...
  • Extension of the DEEP OFFload Programming Model 

    D'Amico, Marco (Universitat Politècnica de Catalunya, 2015-10-21)
    Master thesis
    Restricted access - confidentiality agreement
  • Improving the scheduler of the OmpSs-2 runtime 

    Marí Barceló, Marc (Universitat Politècnica de Catalunya, 2018-06-27)
    Master thesis
    Open Access
    In a task-based model such as OmpSs-2, task creation and scheduling are critical to achieve good performance. For this reason, these areas have been redesigned in the Nanos6 runtime to make them scalable and efficient for ...
  • OmpSs Task offload 

    Sainz Manteca, Florentino (Universitat Politècnica de Catalunya, 2014-07-18)
    Master thesis
    Open Access
    We have extended OmpSs in order to allow users to offload tasks to remote nodes which support MPI (for example, Xeon Phi), while allowing those tasks to communicate internally without having to fall back to the host part ...
  • Rigel: A DSL Use Case for Trace Analysis 

    Aguilar Gallo, Beatriz (Universitat Politècnica de Catalunya, 2015-10-21)
    Master thesis
    Open Access
    Covenantee:  Barcelona Supercomputing Centre
    This Final Master Thesis present a DSL use case to illustrate the potential of DSLs
  • SAIPH: A language for solving PDEs in HPC systems 

    Fernández Suárez, Alejandro (Universitat Politècnica de Catalunya, 2015-06)
    Master thesis
    Restricted access - confidentiality agreement