header team

This website uses cookies to manage authentication, navigation, and other functions. By using our website, you agree that we can place these types of cookies on your device.

View e-Privacy Directive Documents

The Barcelona Supercomputing Center (BSC) was established in 2005 and serves as the Spanish national supercomputing facility. The center hosts MareNostrum4, one of the most powerful supercomputers in the world. The mission of BSC is to research, develop and manage information technologies in order to facilitate scientific progress. BSC strives to be a first-class research centre in supercomputing and in scientific fields that demand high performance computing resources such as Life and Earth Sciences and Computer Applications in Science & Engineering.


DEEP-SEA

BSC will contribute to work package “Co-design application” by providing two DEEP-SEA co-design applications and by providing user support and system administration for small evaluation platforms with innovative technologies that will be tested within the project. In the work package “Measuring, Modelling, Mapping and Monitoring”, BSC will provide a set of tools for application profiling, performance analysis and modelling. BSC’s work in the work package “Software Architecture and System Software” will focus on the memory management. In addition to this, BSC will be involved in the software architecture definition and integration. BSC will lead the work package “Node-level programming”. The main focus of BSC in this work package will be task-based programming models. BSC will also introduce heterogeneous memory awareness to TensorFlow and work on programming environments for processing in memory. In the work package “System-level programming”, BSC will work on malleability, hybrid programming, interoperability and composability and interfaces that support modular/heterogeneous systems. In addition to this, the BSC will take a part of the overall project coordination by taking the roles of:

  • Deputy of the Design and Developers Group (DDG)
  • Coordinator of the Innovation Council

DEEP-EST, DEEP-ER, DEEP

BSC contributed to the DEEP-EST project by leading the work package “Benchmarking and modelling” by experts of the Computer Science Department. The team evaluated the DEEP-EST prototype and measured the benefits that it brings for application end-users and operators of compute infrastructure. BSC’s performance tools and modelling infrastructure (BSC efficiency model) were the basis of this task: Extrae/ Paraver measured the use of compute and memory resources, the interconnect, and the I/O subsystem, while Dimemas provided predictions for architecture or application modifications and the BSC efficiency model yielded accurate scaling predictions (as already demonstrated in the DEEP/-ER projects). BSC tools can automatically identify different phases of applications, and this capability served to build specific models for each of the phases, thus improving precision compared to conventional approaches that treat applications as single unit.

Furthermore, BSC was also involved in the further extension of the "programming environment” that has been developed in the DEEP and DEEP-ER projects to support the heterogeneous MSA and its highly dynamic workloads. In particular, BSC developed its OmpSs programming model further whose offloading features (developed in DEEP and DEEP-ER) were integrated with the resource manager and job scheduler to support efficient execution of evolving and malleable applications on MSA systems.

 

Website: www.bsc.es