HPCwire Soundbite

On today’s episode, we talk with Dr. Michael Heroux of Sandia National Laboratories about the latest benchmark results for the new HPCG rankings. Heroux is a co-founder of the benchmark and is helping to spearhead its development and adoption in the broader HPC community.

As we’ve written about in detail through its development (read here) HPCG is set to be an increasingly common complementary benchmark to Linpack to better represent the real-world needs of applications and systems.

Today we’ll be talking with Dr. Lorena Barba, an Associate Professor of Engineering and Applied Sciences at George Washington University. Barba is a well-known GPU computing proponent and user and is actively involved in reaching broader groups with the CUDA programming message, including most recently female developers. Dr. Barba will describe a recent effort to […]

The post Dr. Lorena Barba Talks “Women Who CUDA” appeared first on HPCwire.

Today we’ll be talking with Dr. Lorena Barba, an Associate Professor of Engineering and Applied Sciences at George Washington University. Barba is a well-known GPU computing proponent and user and is actively involved in reaching broader groups with the CUDA programming message, including most recently female developers. Dr. Barba will describe a recent effort to be featured over the summer at the NVIDIA blog highlighting the work of women CUDA developers, which we will point to when once it’s live. For those ladies out there, it’s not too late to get your voice featured in this series.

Check it out here: http://blogs.nvidia.com/blog/2014/05/08/women-who-cuda-campaign/

We’ll be talking about virtual observatories for big data today with Argonne National Lab’s Dr. Kate Keahey. We’ll discuss the concept of a facility for online data analysis that can support ongoing experiments and other time-critical scientific projects—a facility equipped to deal with millions of sensors, many terabytes of data, and the computational resources required […]

The post Virtual Observatories for Big Scientific Data appeared first on HPCwire.

We’ll be talking about virtual observatories for big data today with Argonne National Lab’s Dr. Kate Keahey. We’ll discuss the concept of a facility for online data analysis that can support ongoing experiments and other time-critical scientific projects—a facility equipped to deal with millions of sensors, many terabytes of data, and the computational resources required to tackle such problems. Kate provides details about what the virtual observatory architecture looks like and offers more background here: http://www.exascale.org/bdec/sites/www.exascale.org.bdec/files/whitepapers/keahey.pdf

The CUDA 6 Performance Report has been released by the engineering team at NVIDIA, and is now available. We’ll talk about how some of the new elements in CUDA 6 are playing out, including the unified memory feature, and talk about real-world performance gains using CUDA 6 with Senior Product Manager for GPU Computing Will […]

The post Real World Performance Gains with CUDA 6 appeared first on HPCwire.

The CUDA 6 Performance Report has been released by the engineering team at NVIDIA, and is now available. We’ll talk about how some of the new elements in CUDA 6 are playing out, including the unified memory feature, and talk about real-world performance gains using CUDA 6 with Senior Product Manager for GPU Computing Will Ramey.