by Calisa Cole

Andrew “Shep” Sheppard recently blogged about a spate of new Meetups cropping up for GPU computing enthusiasts. Since Shep’s post in February, more GPU themed Meetups have launched and many new members have joined. The New York and Silicon Valley groups each have grown from zero to over 100 members in about 10 weeks.

On Monday night, I attended the Silicon Valley GPU Meetup at Carnegie Mellon Silicon Valley, located at the NASA Ames Research Center right next to Moffett Field’s Hangar One.  The meeting was run by group organizer Jike Chong, an adjunct professor at Carnegie Mellon and principal application architect at Parasians, LLC.

Attendees came from a variety of professions – from math aficionados to database experts to a scientist studying the sun. One attendee had just flown in from Singapore and was subsequently awarded a CUDA t-shirt for “longest distance traveled to attend.”

The two-hour meeting gave attendees an opportunity to discuss specific HPC and GPU-related topics like hardware, software, libraries, etc. in small groups, followed by showcase presentations.

Highlights of the evening included:

Application showcase: Jike Chong talked about his work on GPU-accelerated speech recognition, including a description of the challenges. Consider, for example, the nuances between the phrases: “Recognize Speech” and “Wreck a Nice Beach” and “Reckon Eyes Peach.” (Learn more:

Infrastructure showcase: Tim Child spoke on the topic of “Database Driven OpenCL Programming.” Potential usage cases include 3D content management and DNA sequence matching. (Learn more:

Application development: Paulius Micikevicius of NVIDIA gave an overview of Visual Profiler and described new features such as analysis windows and new counters for memory access. (Learn more:

These events are great ways to connect with locals with similar interest in HPC and GPU computing.  Maybe there’s a group in your area?