High Performance Computing

Build scalable GPU-accelerated applications. Faster.

Researchers, scientists, and developers can accelerate their High Performance Computing (HPC) applications using specialized libraries, directive-based approaches, and language-based models. CUDA-X HPC, OpenACC, and CUDA help developers utilize the thousands of computational cores available on NVIDIA GPUs to deliver ground breaking application performance in domains ranging from computational science to artificial intelligence. The applications can be developed, optimized and deployed using popular languages such as C, C++, Python, Fortran, and MATLAB.

  • CUDA-X HPC, a collection of GPU-accelerated libraries built on CUDA, provide the fastest path to accelerate a wide range of HPC applications through highly-optimized drop-in functions.
  • OpenACC is a directive-based programming model designed to help scientists and researchers accelerate their codes with significantly less programming effort than required with a low-level model.
  • CUDA®, NVIDIA’s parallel computing platform and programming model, offers a language-based solution for programmers who want to fine tune their applications for the best possible performance.

These complementary solutions are applied either separately or in concert to accelerate applications on desktops, workstations, enterprise and hyperscale data centers, and the fastest supercomputers on the planet. The choice of solution is based on a combination of factors such as type of code, desired performance gains, and programming effort.


CUDA Toolkit
Complete environment for building scalable GPU-accelerated applications in C, C++, Fortran and Python.

GPU-Accelerated Libraries
CUDA-X HPC, a collection of optimized GPU-accelerated libraries for commonly used computing operations, included in both the CUDA and OpenACC Toolkits.

OpenACC Toolkit
Directive-based solution providing simple yet powerful approach to accelerators.


Designed to power energy efficient applications that will be deployed on everything from the largest enterprise data centers and supercomputers with thousands of GPUs to resource-constrained embedded devices.

Wide Availability

Gives you maximum flexibility when choosing GPU-accelerated platforms for development and deployment.

Learn More

Complete Ecosystem

Integrated with commonly-used programming languages, top numerical packages and deep learning frameworks.

HPC Everywhere

>70% HPC Apps accelerated


19,000+ organizations engaged

Get Started With Hands-On Training

The NVIDIA Deep Learning Institute (DLI) offers hands-on training for developers, data scientists, and researchers in AI and accelerated computing. Get started online with hands-on, self-paced courses on the fundamentals of CUDA C/C++, CUDA Python, and OpenACC today.