The GPGPU Continuum

  • Date/Time
    Date(s) - 25/04/2013
    8:30 am - 4:30 pm

    Location
    Room 1003, EE Meyer Building

    Categories No Categories


    TCESagivNvidialogos
     
     
     
     
     
     
     
    The GPGPU Continuum from mWatts to peta flops
     
    Organizing committee: Prof. Avi Mendelson, Ofer Rosenberg, Dr. Chen Sagiv
    Date: Thursday, April 25th 2013
    Location: Room 1003, EE Meyer Building, Technion, Haifa
    Time: 08:30-16:30
     
    Participation is free but requires registration. Please register HERE
     
    The GPU continuum workshop addresses the current and future challenges of the GPGPU (heterogeneous systems) community. GPUs that started as graphics accelerators are now widely used in many different computational domains, such as medical, defense, silicon inspection, computer vision, signal processing and more, due to their dramatic improvement of the performance per power for domain specific applications.
     
    This workshop aims to address two important directions of future GPGPU based systems: at the high-end, it will focus on using GPUs as part of clusters, clouds, compute farms, etc., that aim to achieve peta flops of computations. At the low-end, it will focus on using GPUs as part of mobile devices, which limits the power consumption of the GPU to mWatts. We conclude the workshop with a panel discussion on the differences and similarities and on the challenges each domain faces.
     
     
    Slides and vidoes now available below
     
    Program
     
    Time
    Topic
    Speaker
    Video/Slides
    8:30 – 9:00
    Coffee & get together
     
     
    9:00 – 9:10 Welcome Prof. Oded Shmueli VP Technology Technion  Video
    9:10 – 9:20
    Welcome and introduction to the Technion new HPC center
    Prof. Avi Mendelson, Technion
     Video
    9:20-10:00
    Keynote
    Guillermo (Eli) Savransky, Principal Architect, CTO Office, Mobile Business Unit Nvidia Corp
     Video
    10:00-10:30
    MultiGPU infrastructure for real time applications
    Eri Rubin, SagivTech
     Video
    10:30-11:00
    Coffee Break sponsored by NVIDIA
     
    11:00-11:30
    New Models for GPU to GPU Communications
    Gilad Shainer, Mellanox
     Video | Slides
    11:30-12:00
    GPUfs: integrating a file system with GPUs
    Dr. Mark Silberstein – TU Austin & Technion
     Video | Slides
    12:00-12:30
    ClusterCL: Extending OpenCL with cluster support
    Prof. Amnon Barak & Jonathan Smith, HUJI
     Video
    12:30 – 13:00
    OpenCL for Intel Xeon Phi: Internals and Optimizations
    Dr. Yariv Aridor, Intel
     Video Slides
    13:00-14:00
    Lunch sponsored by NVIDIA
    14:00-14:20
    The GPGPU continuum
    Ofer Rosenberg
     Video Slides
    14:20-14:50
    Energy Efficient Heterogeneous Computing for a New Mobile Experience
    Rotem Aviv, Qualcomm
     Video | Slides
    14:50-15:20
    Taking GPGPU to the ULT space –Intel Haswell Architecture
    Michael Behar, Intel
    15:20-15:50
    Heterogeneous computing architecture for low power devices
    Tzahi Cohen, AMD
     Video | Slides
    15:50-16:10
    Coffee Break sponsored by NVIDIA
    16:10-16:40
    Accelerating Ultrasound imaging using GPGPU – overview and challenges
    Tomer Gal, GE
     N/A
    16:40-17:10
    HPC based activities as part of TCE – Technion
    Mr. Uri Verner
     Video
    17:10-
    Summary and Open Discussion
    Prof.  Avi Mendelson, Technion