by Chandra Cheij

Researchers from Tokyo Institute of Technology snagged the first-ever Achievement Award for CUDA Centers of Excellence (CCOE), for their research with TSUBAME 2.0.

The team was among three other groups of researchers from CCOE institutions, which include some of the world’s top universities engaged in cutting-edge work with CUDA and GPU computing.

Each of the world’s 18 CCOEs was asked to submit an abstract describing their top achievement in GPU computing over the past year and a half. A panel of experts, led by NVIDIA Chief Scientist Bill Dally, selected four CCOEs to present their achievements at a special event during GTC 2012 this week in San Jose. CCOE peers voted for their favorite, who won bragging rights as the inaugural recipient of the CUDA Achievement Award 2012.

The four finalists – each of whom received an HP ProLiant SL250 Gen8 system configured with dual NVIDIA Tesla K10 GPU accelerators – are described below. Abstracts of their work are available on the CCOE Achievement Award website.

Barcelona Supercomputing Center, OmpSs: Leveraging CUDA for Productive Programming in Clusters of Multi-GPU Systems

OmpSs is a directive-based model through which a programmer defines tasks in an otherwise sequential program. Directionality annotations describe the data access pattern for the tasks and convey the runtime information it uses to automatically detect potential parallelism, to automatically perform data transfers and to optimize locality. Integrating this model with CUDA allows applications to leverage the dazzling performance of GPUs, enabling the same simple and clean code that would run on an SMP to run on multi-GPU nodes and clusters.

Harvard University, Massive Cross-Correlation in Radio Astronomy with Graphics Processing Units

The study of the universe is no easy task. Rather than struggle to build larger and larger telescopes in their challenge to understand our vast universe, Harvard University is using GPU computing technologies to help them create telescope arrays composed of many smaller telescopes. Harvard researchers have developed the Harvard X-Engine code to help integrate data from these types of telescope arrays, with an emphasis on removing data-crunching bottlenecks.

Tokyo Tech, TSUBAME 2.0

Researchers at the Tokyo Institute of Technology have designed and constructed Japan’s first petascale supercomputer, known as TSUBAME 2.0, as well as a series of advanced software and research applications. Such activities have been rewarded with numerous results presented at top academic venues as well as numerous global accolades and press reports. Tokyo Tech highlighted the three core achievements of TSUBAME / CUDA CCOE, but the results are not just limited to them.

University of Tennessee, MAGMA: A Breakthrough in Solvers for Eigenvalue Problems

Scientific computing applications – ranging from those that help analyze how earthquakes propagate through a medium and affect bridges, to those that simulate energy levels of electrons in nanostructure materials – require the solution of eigenvalue problems. The Matrix Algebra on GPU and Multicore Architectures (MAGMA) project aims to develop algorithms that will speed up computations on heterogeneous multicore-GPU systems by at least one order of magnitude.