With the GPU Technology Conference less than a month away, I took a look through the sessions catalog today and realized I’m going to have a hard time deciding what to attend – there are more than a dozen tutorials and over 250 sessions covering everything from algorithms to visualization.

So I’ve put together a top 10+ list of the sessions I’m most looking forward to attending.

As a GPU Computing product manager and former software engineer, I’m interested in a wide breadth of topics, but I’ll be sprending most of my time at GTC between sessions on developer tools, programming techniques and research topics. If those are your interests, too, this list will be right up your alley.

If you want to catch meI’ll also be presenting a pre-conference tutorial on, Languages, APIs and Development Tools for GPU Computing, Monday, September 20th, 1:00 – 2:20pm. And if you still haven’t registered, make sure you do so before Sept. 1 to take advantage of the early bird pricing.

Without further ado, here are my top 10+3 (in no particular order):

  • Analysis-Driven Performance Optimization (2012):
    This 2hr session was one of the most popular last year, showing how to use simple tools and techniques to identify and crush performance bottlenecks.
  • Analyzing CUDA Accelerated Application Performance at 20 PFLOP/s (2089):
    Presented by the team that develops VampirTrace, these guys know how to analyze petascale performance.
  • Any session on debugging:
    There are several sessions on best practices for debugging, including a basic session on CUDA Debugging on Linux (2002), GPU Debugging with Allinea DDT (2039), and TotalView Debugger for CUDA (2251)
  • Overview of Parallel Nsight for Visual Studio (2149):
    This is a good chance to get a closer look at the integrated development, debugging and performance analysis features in the latest release of NVIDIA’s Parallel Nsight for Visual Studio.
  • Integrating CUDA into a Large-Scale Commercial Database System (2092):
    These guys have figured out how to hook up SQL Server to GPUs and are processing terabytes of data. Very cool!
  • GPU Computing with MATLAB (2267):
    I’m really looking forward to seeing how easy it is to take advantage of GPU acceleration from within MATLAB.
  • CUDA Libraries Open House (2216:
    This is an opportunity to learn about the latest features in NVIDIA’s math libraries and meet the team working on them.
  • CULA – A Hybrid GPU Linear Algebra Package (2153):
    Doesn’t everyone need highly optimized linear algebra libraries for heterogeneous systems?
  • Sessions on Statistical Analysis using R with GPU Acceleration: I’m planning to attend both Using R for High-Performance Data Analysis (2111) and An R Library for Native GPU Objects (2179) to learn more about how people using the open source R package are benefiting from GPU acceleration.
  • High-Productivity CUDA Development with the Thrust Template Library (2219):
    Because higher productivity is something everyone needs!
  • BCSLIB-GPU: Significant Performance Gains for CAE (2213):
    I’m really looking forward to seeing the results this blazing fast port of the Gordon Bell Award winning library of sparse matrix routines can achieve on GPUs.
  • Accelerating Signal Processing: Introduction to GPU VSIPL (2126):
    We’ve known for a while that GPUs are really good at signal processing. Now there’s a really easy way to take advantage of GPU-accelerated signal processing in just about any application.
  • The Best of Both Worlds: Flexible Data Structures for Heterogeneous Computing (2038):
    The speaker will be presenting some simple techniques for defining data structures that are optimal for use on both CPUs and GPUs, including some AoS / SoA magic.
  • Languages, APIs, and Development Tools for GPU Computing (2004):
    This is the tutorial I’m presenting on Monday, so I can’t miss this one! You shouldn’t either if you want to get your arms around the basics of GPU Computing before attending the more advanced sessions later in the conference.

There are also a bunch of interesting sessions on GPU Algorithms, Finance, Film, Medical Imaging, Manufacturing/CAE, Energy Exploration, Life Sciences and more… but I’ll let you browse through those on your own. The full sessions catalog is available here – you should book your seat in advance using the scheduler, for the sessions you're planning to attend to ensure you'll have a slot.  And, if you’d like to share your own Top 10 GTC sessions or ask questions about the conference, please post them in the comments below. Hope to see you in San Jose on September 20th!

Similar Stories