8th Workshop on General Purpose Processing Using GPUs

Overview

The final program is available HERE

The goal of this workshop is to provide a forum to discuss new and emerging general-purpose purpose programming environments and platforms, as well as evaluate applications that have been able to harness the horsepower provided by these platforms. This year's work is particularly interested on new heterogeneous GPU platforms, new forms of concurrency, and novel/irregular applications that can leverage these platforms. Papers are being sought on many aspects of GPUs, including (but not limited to):

    • GPU applications
    • GPU programming environments
    • GPU runtime systems
    • GPU compilation
    • GPU architectures
    • Multi-GPU systems
    • GPU power/efficiency
    • GPU reliability
    • GPU benchmarking/measurements
    • Heterogeneous GPU platforms that incorporate GPUs

    • Keynote

    Dark secrets of heterogeneous memory models

    Lee Howes, Qualcomm

    Abstract:

    As GPU programming has become more advanced, heterogeneous systems more common and programming approaches more flexible, developers have wanted more control over memory in the system. In recent specifications, OpenCL 2.0 and the first HSA documentation that has been released, we see a stronger focus on memory consistency. Not all is as clear cut as it seems and there is still work to do. In this talk we'll look at what has changed recently, where some of the gaps are and what we might like to do to improve the situation.

    Biography:

    Lee Howes works on GPU architecture and programming models at Qualcomm. He represents Qualcomm at both Khronos, for OpenCL and the memory model, execution model and intermediate language working groups, as well as to the HSA Foundation on the runtime and intermediate language specifications. Lee previously worked at AMD in the HSA organization and before that on physics simulation on heterogeneous platforms. He has a PhD in software performance optimization from Imperial College London.

    Registration

    Please refer to the PPoPP 2015 website for registration and local information.

    Workshop Organizers

    David Kaeli, Northeastern University
    John Cavazos, University of Delaware

    Program Committee

    Jose Luis Abellan - UCAM
    Sonia Alarcon - Rochester Institute of Technology
    Brad Beckmann - AMD
    Ramon Canal - Polytechnic University of Catalonia
    Sunita Chandrasekaran - University of Houston
    Rodrigo Dominguez - Qualcomm
    Magnus Ekman - NVIDIA
    Isaac Gelado - NVIDIA
    Lee Howes - Qualcomm
    Byunghyun Jang - University of Mississippi
    Ajay Joshi - Boston University
    Paul Kelly - Imperial College
    James Lin - Shanghai Jiao Tong University
    Brian Lewis - Intel
    Simon McIntosh-Smith - University of Bristol
    Richard Membarth - Saarland University
    Avi Mendelsohn - Technion
    Perhaad Mistry - AMD
    Dimitris Nikolopoulos - Queen's University Belfast
    Oscar Plata - University of Malaga
    Norm Rubin - NVIDIA
    Weibin Sun - Google
    David Wood - University of Wisconsin-Madison
    Huiyang Zhou - North Carolina State University