The SO Seminar is a usually monthly event developed in the period from September to July and organised with the support of FIB-Barcelona School of Informatics, UPC BarcelonaTech.

Sumari

  • Exploring emerging technologies for extreme scale HPC architectures 

    Vetter, Jeffery S. (2016-09-10)
    Text en actes de congrés
    Accés obert
    While architectures and programming models have remained relatively stable for almost two decades, new architectural features, such as heterogeneous processing, nonvolatile memory, and optical interconnection networks, ...
  • Kernel and user level execution trace analysis for multi-core distributed Linux systems 

    Dagenais, Michel (2016-09-10)
    Accés obert
    The presentation will briefly introduce the organization of the DORSAL laboratory and the collaborative research and development projects undertaken with the financial support of industrial partners, including Ericsson, ...
  • Learning by redundancy: climate multi-model ensembles and machine learning 

    De Felice, Matteo (2016-09-10)
    Text en actes de congrés
    Accés obert
    Climate Models are sophisticate tools able to simulate the interactions among various components of the Earth system (atmosphere, oceans, bio-sphere, etc.). Those tools are nowadays used for many purposes: to improve the ...
  • Datacenter computers: modern challenges in CPU design 

    Sites, Richard L. (2016-09-10)
    Text en actes de congrés
    Accés obert
    Computers used as datacenter servers have usage patterns that differ substantially from those of desktop or laptop computers. We discuss four key differences in usage and their first-order implications for designing computers ...
  • Simulating extreme hydro-meteorological events with DRIHM(2US) services 

    Parodi, Antonio (2016-09-10)
    Text en actes de congrés
    Accés obert
    From 1970 to 2012, about 9000 High Impact Weather Events (HIWE) were reported globally: all together, they caused the loss of 1.94 million lives and economic damage of US$ 2.4 trillion (2014 UNISDR report). Storms and ...
  • A New role of ontologies and advanced scientific visualization in big data analytics 

    Chuprina, Svetlana (2016-09-10)
    Text en actes de congrés
    Accés obert
    Accessing and contextual semantic searching structured, semi-structured and unstructured information resources and their ontology based analysis in a uniform way across text-free Big Data query implementation is a main ...
  • The Route towards the ultimate network topology 

    Rodriguez, German (2016-09-10)
    Text en actes de congrés
    Accés obert
    In this talk I will try to summarize our quest for a realizable network topology that optimizes performance, cost, power consumption and partitionability. We have explored Fat Trees, Dragonflies, variations of dragonflies, ...
  • Adaptive, efficient, parallel execution of parallel programs 

    Sohi, Guri (2016-09-10)
    Text en actes de congrés
    Accés obert
    Future parallel processors will be heterogeneous, be increasingly less reliable, and operate in dynamically changing operating conditions. This will result in a constantly varying pool of hardware resources which can greatly ...
  • VIPS: simple, efficient, and scalable cache coherence 

    Ros, Alberto (2016-09-10)
    Report de recerca
    Accés obert
    Directory-based cache coherence is the de-facto standard for scalable shared-memory multi/many-cores and significant effort is invested in reducing its overhead. However, directory area and complexity optimizations are ...
  • Towards real-time CFD simulation of in-flight icing 

    Habashi, Wagdi G. (2016-09-10)
    Text en actes de congrés
    Accés obert
    Despite the concerted efforts of manufacturers and certification agencies, incidents and accidents continue to happen to aircraft certificated to “Fly Into Known Icing”, falsely thought of as being “immune to in-flight ...
  • Semantically ordered parallel execution of multiprocessor programs 

    Sohi, Guri (2016-09-10)
    Text en actes de congrés
    Accés obert
    Nondeterministic execution of conventional, multithreaded programs complicates multiprocessor programming and system use in general. To eliminate nondeterminism from multiprocessor programs, we propose an ordered approach. ...
  • The Fourth paradigm - data-intensive scientific discovery and open science 

    Hey, Tony (2016-09-10)
    Text en actes de congrés
    Accés obert
    There is broad recognition within the scientific community that the ongoing deluge of scientific data is fundamentally transforming almost all areas of academic research. A wide variety of researchers —from scientists and ...
  • Advances in GPU architecture for deep learning and scientific computing 

    Parienté, Frédéric (2016-09-10)
    Text en actes de congrés
    Accés obert
    The talk will cover the recent NVIDIA product announcements made at the GTC'16 conference, and how the Pascal GPU and NVLink interconnect technologies greatly improve multi-GPU performance and efficiency in deep learning ...
  • Semantic systems and visual tools to analyze climate change communication 

    Scharl, Arno (2016-09-10)
    Report de recerca
    Accés obert
    Given the intense attention that environmental topics such as climate change attract in news and social media coverage, key questions are how different stakeholders perceive observable threats and policy options, how public ...
  • Talk 1: Convolutional neural networks against the curse of dimensionality 

    Bruna, Joan (2016-09-10)
    Text en actes de congrés
    Accés obert
    Convolutional Neural Networks are a powerful class of non-linear representations that have shown through numerous supervised learning tasks their ability to extract rich information from images, speech and text, with ...
  • Talk 2: Recurrent neural nets and differentiable memory mechanism 

    Vinyals, Oriol (2016-09-10)
    Text en actes de congrés
    Accés obert
    This past year, RNNs have seen a lot of attention as powerful models that are able to decode sequences from signals. The key component of such methods are the use of a recurrent neural network architecture that is trained ...
  • Large-scale machine learning in cancer and brain research: new applications that will drive future supercomputing systems 

    Stevens, Rick (2016-09-10)
    Text en actes de congrés
    Accés obert
    In this talk I’ll discuss the DOE/NCI cancer research co-design projects that are a key part of the Presidential Moonshot Cancer Initiative and the brain connectome project at the heart of the National Brain Observatory ...