At SC26, NVIDIA is marking the 20th anniversary of CUDA, the platform that helped establish GPU computing as a foundation for modern science and artificial intelligence. The anniversary recognizes the researchers, developers and engineers whose experimentation, tools and persistence expanded accelerated computing from graphics into scientific simulation, deep learning and other demanding workloads.
CUDA’s history also reflects the evolution of computing over the past two decades—from the first developers writing CUDA programs to today’s AI researchers building on that work. More than 11 million developers now use the platform, with the broader computing stack supported by CUDA.
https://www.youtube.com/watch?v=v2I4vqz87fk" title="20 years of CUDA
From SC06 to SC26
NVIDIA announced CUDA in November 2006 and introduced it to the scientific community at SC06. The platform gave scientists and developers a familiar, high-level programming model for using GPUs, extending accelerated computing beyond graphics to applications such as science and simulation.
Over the 20 years since then, developers have used CUDA across a wide range of fields. The platform has supported AI research associated with a Nobel Prize, accelerated scientific applications on supercomputers and powered visual effects for Oscar-nominated films.
CUDA’s reach has grown alongside advances in computing. Its developer community and software ecosystem have helped make GPUs a central tool for research, engineering and artificial intelligence.
Celebrating the architects of progress
At SC26, NVIDIA is highlighting several CUDA pioneers whose work helped advance science and AI.
Molecular simulation and the HIV capsid
At the University of Illinois at Urbana-Champaign, biophysicist Klaus Schulten and his team helped connect GPU computing with molecular biology through nanoscale molecular dynamics and Visual Molecular Dynamics, a tool for molecular simulation and visualization.
In 2013, researchers from the University of Illinois, the University of Pittsburgh and Vanderbilt University used CUDA-accelerated tools to help determine the atomic structure of the HIV capsid. The team combined experiments and simulations on Blue Waters, which was one of the world’s most powerful supercomputers at the time.
GPU computing and the rise of deep learning
University of Toronto researchers Alex Krizhevsky, Ilya Sutskever and Geoffrey Hinton used CUDA in the development of AlexNet. Running on just two NVIDIA GPUs, their image-recognition system won the 2012 ImageNet competition and helped demonstrate the potential of GPU computing for deep learning.
AlphaFold and protein structure prediction
The link between computing and scientific discovery is also evident in AlphaFold, which predicts protein structures. Demis Hassabis and John Jumper jointly won the 2024 Nobel Prize in Chemistry for work on protein structure prediction. Their team’s AlphaFold 2 software includes CUDA support for GPU execution.
These examples represent only a small part of the ways one generation’s work has created opportunities for the next. With more than 1,000 NVIDIA CUDA-X libraries available, developers are applying accelerated computing to areas including drug discovery, next-generation factory design, autonomous vehicles, financial-market modeling and quantum computing.
Developers can start building with the CUDA Toolkit and visit the 20 Years of CUDA Celebration Event Wall at the NVIDIA booth at SC26.
Source: blogs.nvidia.com


