- NVIDIA DGX-1
- NVIDIA GRID
- Quadro VCA
- NVIDIA DRIVE
- About NVIDIA
Our GPUs make enormous computational power available to a wider number of people than ever before. They enable scientists, researchers and innovative companies to accomplish projects that were once out of reach because the expense and difficulty of obtaining supercomputing resources.
The parallel processing capability of the GPU allows it to divide complex computing tasks into thousands of smaller tasks that can run concurrently.
As a result, computational scientists and researchers are seeing results in days instead of months, even minutes instead of days. GPUs enable them to address some of the world’s most challenging computational problems up to several orders of magnitude faster than before – whether their work is focused on the relationship between protein folding and human disease or on the search for clues along the outer edge of our galaxy.
At the University of Antwerp, Belgium, a team uses an eight-GPU system to perform large-scale scientific computations in its research on tomography, a medical imaging technique that reconstructs large images from X-rays into 3D views. The FASTRA GPU SuperPC is able to perform the team’s computations just as fast as a cluster containing 350 CPU cores. It does this in a desktop footprint, without the expense, power consumption or maintenance of a large CPU cluster.
Watch the video here: http://www.youtube.com/embed/_l6ZjquXG2Q
NVIDIA’s Tesla technology makes high-performance computing accessible at lower power and cost than traditional CPU-based supercomputing.
This puts supercomputer performance within easier reach for students and researchers who previously may not have been able to access it. Particular beneficiaries are scientists in countries without national computing facilities who can now participate in research that previously only developed nations could pursue.
Through NVIDIA Research, we work with researchers worldwide who are interested in using the GPU to increase their computational power and reduce time to discovery.
Massively parallel NVIDIA GPUs have also enabled supercomputing performance in a personal computer. By adding Tesla GPU processors to personal workstations, our customers can have cluster-level computing performance—up to 250 times faster than standard PCs and workstations—right on their desks.
“[T]he Tesla products essentially give us all a personal supercomputer. Just one Tesla C1060 delivers the same performance as our 64 CPU cluster, and this was a resource we had to share. This is a huge cost and time saving that has transformed our workflow and boosted our productivity.”
--James Allison, President, OpenGeoSolutions, Inc.
“We allow scientists who are solving incredibly hard problems to move much faster.”
--Andy Keane, General Manager, Tesla High-Performance Computing Solutions
"In January 2010, I attended the NVIDIA-sponsored "Accelerated Computing" conference, organized by the Riken Institute. The organizers invited me to speak about the technical aspects of the research in my group using GPU computing. Instead, I decided to speak about GPU as a technology that can level the playing field for scientists in many parts of the world to participate in leading-edge computational research. I referred to "scientifically developing countries", where they have very good education and many eager bright young minds, but cannot afford big investments in research infrastructure.”
--Lorena Barba, Assistant Professor, Department of Mechanical Engineering, Boston University
The Instituto Tecnologico de Estudios Superiores de Monterrey (ITESM) uses GPU computing with NVIDIA’s CUDA architecture to solve research problems in large-scale, real-time computer graphics that students at the university could not do before.
Students have been able to complete PhD dissertations by using GPU computing to study graphics problems such as crowd simulations and generation of diversity for crowds, generating and animating thousands of complex characters from a few representative models.
In other, non-graphics uses, students have applied neural nets and genetic algorithms to technical indicators for financial predictions using the GPU and achieved a 20X speedup for these calculations using CUDA. ITESM is teaching parallelism in a graduate-level course and pursuing further research in graphics.
-Isaac Rudomin, Ph.D. (UPenn), Associate Professor, ITESM-CEM