CUDA: From HIV Simulations to AI Models

Molecular biology and image recognition are among the fields where CUDA has been used to leverage GPUs. NVIDIA's platform is also utilized for the visual effects of Oscar-winning films.
Biologists Harness CUDA to Study Molecules
In 2013, scientists from Illinois, the University of Pittsburgh, and Vanderbilt University contributed to determining the atomic structure of the HIV capsid using tools that employ CUDA. Their work combined experiments with simulations conducted on Blue Waters, which was then among the most powerful supercomputers in the world. At the University of Illinois at Urbana-Champaign, biophysicist Klaus Schulten's team had facilitated the connection between GPUs and molecular biology with Nanoscale Molecular Dynamics and Visual Molecular Dynamics, designed for the simulation and visualization of molecules.
CUDA is also involved in AlphaFold 2, software made public by Demis Hassabis and John Jumper's team: its execution on GPUs relies on the platform. AlphaFold is used to predict protein structures. Hassabis and Jumper shared the 2024 Nobel Prize in Chemistry for their work on this prediction.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
AlexNet Won ImageNet with Training on Two GPUs
At the University of Toronto, Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton trained AlexNet with CUDA on just two NVIDIA GPUs. This image recognition system won the ImageNet competition in 2012. This result demonstrated the potential of GPUs for deep learning.
The NVIDIA CUDA-X offering now brings together more than 1,000 libraries, used notably for developing machine vision, learning, and reasoning. Its areas of application include drug discovery, next-generation factory design, autonomous vehicles, financial modeling, and quantum computing.
NVIDIA Gathers Pioneers for CUDA's 20th Anniversary
More than 11 million developers now rely on CUDA at various levels of the computing stack. NVIDIA is celebrating its 20th anniversary this week at SC26, where the company honors pioneers as well as researchers, engineers, and developers who have contributed to scientific uses and AI. The tools and experiments of this community have expanded the possibilities of accelerated computing; today's AI developers build upon the work of the platform's early programmers.
Announced in November 2006 and then presented to scientists at SC06, CUDA offered them a familiar high-level programming model to leverage GPUs. Its use for accelerated computing has thus extended beyond graphics, particularly into science and simulation.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.