Proven Algorithmic Techniques for Many-core Processors


Proven Algorithmic Techniques for Many-core Processors
August 2–6, 2010
Enroll online now!
Enroll for the 2010 summer school courses at:



For a fee of $100, students can participate at state-of-the-art classrooms across the country. The distributed sites are linked by high-definition videoconferencing, enabling synchronous audio/video Q&A. Students can benefit from hands-on lab activities with on-site support from skilled teaching assistants. All students will have access to high-performance computing resources during the course and will receive certificates of completion.
Snacks and an evening reception will be provided for on-site students, but participants are responsible for their own travel and lodging costs (low-cost dorm accommodations will be provided where possible). Students are required to provide their own laptops.
For no additional cost, on-site participants can take an online short course on CUDA that are designed to help them meet course prerequisites. These courses are available only to on-site participants.
NOTE: On-site participation is "first come, first served" and seating at each site is limited. Register early to reserve a seat at your preferred site!
Students who register in April will be notified of their site assignment by May 7. Those who register in May or after will be notified within two weeks of registration.
Once you are notified of your site assignment, you have two weeks to pay the course fee and complete the registration process before you forfeit your spot.
Center for Computation & Technology, Louisiana State University, Baton Rouge
Institute for Data and High Performance Computing, Georgia Institute of Technology, Atlanta
Institute for Digital Research and Education, University of California, Los Angeles
National Center for Supercomputing Applications, Urbana, Illinois
Northwestern University, Evanston, Illinois
Ohio Supercomputer Center, Ohio State University, Columbus
RENCI, Chapel Hill, North Carolina
University of Iowa, Iowa City
University of Michigan, Ann Arbor
University of Tennessee, Knoxville
  • Experience working in a Unix environment
  • Experience developing and running scientific codes written in C or C++
  • Basic knowledge of CUDA (A short online course, Introduction to CUDA, is available to registered on-site students who need assistance in meeting this prerequisite)
Students who took the course Many-core Processors in 2009 are encouraged to take this follow-on course, which includes new topics and lab exercises.
  • Wen-Mei W. Hwu, professor of electrical and computer engineering and principal investigator of the CUDA Center of Excellence, University of Illinois at Urbana-Champaign
  • David Kirk, NVIDIA fellow
Course outline:
  • Introduction
    • why problem formulation and algorithm design choices can have dramatic effect on performance
    • common algorithmic strategies for high performance
  • Increasing locality in dense arrays
    • tiling of data access and layout
  • Improving efficiency and vectorization in dense arrays
    • granularity coarsening
  • Reducing output interference
    • conversion from scatter to gather
    • parallelizing reductions and histograms
  • Dealing with non-uniform data
    • data sorting and binning
  • Dealing with sparse data
    • sorting and packing
  • Dealing with dynamic data
    • parallel queue-based algorithms
  • Improving data efficiency in large data traversal
    • stencil and other grid-based computation
  • Extending beyond many-core processors
    • MPI+CUDA
    • MPI+OpenCL
  • Overview of use of techniques in application domains
    • molecular dynamics
    • computational fluid dynamics
    • medical imaging
    • computer vision
    • gene sequencing
  • Case studies:
    • molecular dynamics (NAMD/VMD, MPI, use of algorithm strategies)
    • medical imaging
    • gene sequencing, financial analysis, etc.
  • Hands-on Lab
NOTE: Students are required to provide their own laptops.


Event Date & Time: 
Monday, August 15, 2011 - 00:00 to Friday, August 19, 2011 - 00:00