Faster planes. More fuel-efficient cars. Wind tunnels are incredibly useful tools. Even high-performance clothes used by runners and bikers are tested in them. But a world-class wind tunnel isn’t cheap, and time in one is precious. One solution: virtual wind tunnels.
Building on the work to create special-purpose software using algorithms well-suited to GPU computing, students at Germany’s TU Clausthal crafted a virtual wind tunnel that uses an NVIDIA GeForce GTX 690 GPU to simulate airflow on a computational grid with 300,000 cells. The result: a smooth visualization of the transient flow of air over an object. All done on a GPU that can be picked up at the nearest electronics store.
TU Clausthal was among the more than three dozen institutions from 11 countries that were added this past quarter to our roster of CUDA Research Centers and CUDA Teaching Centers. There are now a total of 293 CUDA Research and Teaching centers in 43 countries.
CUDA Teaching Centers equip tens of thousands of students graduating each year with the knowledge and expertise to take advantage of the parallel processing power of GPUs (see “What Is CUDA?”). They get free teaching kits, textbooks, software licenses, NVIDIA CUDA architecture-enabled GPUs for teaching lab computers and academic discounts for additional hardware.
CUDA Research Centers embrace GPU computing across multiple research fields. They have access to exclusive events with key researchers and academics, a designated NVIDIA technical liaison and specialized training sessions.
Here are a few more examples of CUDA-related work taking place at some of our newest CUDA Research Centers:
Providing a catalyst for discussion between NVIDIA and the Illinois Institute of Technology (IIT) and their collaborators, the University of Chicago, Argonne National Laboratory and Cray, the CUDA Research Center at IIT was established to support a variety of GPGPU research. The core research is centered on making GPUs a viable computing platform for Many-Task Computing (MTC) applications. This work aims to address the programmability gap between the MTC paradigm and many-core accelerators through an innovative CUDA middleware GeMTC (GPU enabled Many-Task Computing) coupled with the Swift implicitly parallel data-flow driven programming system.
RIKEN (the Advanced Center for Computing and Communication) is Japan’s largest and most comprehensive research organization for basic and applied science and a world leader in a diverse array of scientific disciplines. The CUDA Research Center at RIKEN provides its researchers a computing environment and opportunity for consultation to tune their program codes.
Researchers at the Vishwakarma Institute of Technology in Pune are investigating Fuzzy Hyper Line Segment Neural Networks (FHLSNN). The networks learn patterns in terms of hyper lines, which are fuzzy sets and are associated with their fuzzy membership function. The institute plans to speed up training and testing phases of FHLSNN on the NVIDIA Tesla K20 GPU for high-dimensional, large data sets.
Additional new CUDA Research Centers include:
- Centro Extremeño de Tecnologías Avanzadas (Spain)
- Indian Institute of Technology, Kanpur (India)
- Wuhan University (China)
The new CUDA Teaching Centers include:
- City College of San Francisco (U.S.)
- Inner Mongolia University (China)
- New York University (U.S.)
- Northwestern Polytechnical University (China)
- Rajarambapu Institute of Technology (India)
- RK University (India)
- Shanghai University (China)
- Tokyo Metropolitan University (Japan)
- TU Delft (Netherlands)
- Universidad Nacional de Cuyo (Argentina)
- Universidade Federal Rural de Pernambuco (Brazil)
- Universitat Autònoma de Barcelona (Spain)
- University of Calcutta (India)
For more information on NVIDIA research activities and these programs, visit the NVIDIA Research site.