CUDA-Q Platform

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

type access

  • Operating System:

  • Terminal:

  • Shell:

  • Editor:

  • Package Manager:

  • Programming Language:

  • Database:

  • NVIDIA Libraries:

  • Extension:

NVIDIA CUDA-Q is a platform for hybrid quantum-classical computing. It provides a unified programming model that enables CPUs, GPUs, and QPUs (quantum processing units, either physical or emulated) to work together within a single workflow.

At the core of CUDA-Q is NVQ++, a compiler that implements the CUDA-Q specification and enables execution across different backends. Information on supported simulation backends is available here.

Initialization

For information on how to use the Initialization parameter, refer to the Initialization: Bash Script and Initialization: PyPI Packages sections of the documentation.

Run a Script in Batch Mode

CUDA-Q supports a batch mode option for jobs. When enabled, a .sh script runs once at job start, and the job ends as soon as the script finishes execution. This is useful for automating workflows, quick experiments, or launching pre-defined workloads.

Examples and Applications

CUDA-Q includes a collection of examples and applications located in the default user’s home directory:

  • Examples – showcase core programming concepts.

  • Applications – demonstrate real-world use cases.

Example: Bernstein–Vazirani algorithm

The following command compiles and executes an implementation of the Bernstein-Vazirani algorithm:

$ nvq++ $HOME/examples/cpp/basics/bernstein_vazirani.cpp && ./a.out

If a suitable GPU is detected, CUDA-Q automatically applies GPU acceleration. To confirm GPU usage, the problem size can be increased by modifying the secret string length and explicitly targeting the NVIDIA backend:

$ nvq++ $HOME/examples/cpp/basics/bernstein_vazirani.cpp -DSIZE=25 --target nvidia && ./a.out

This configuration typically demonstrates improved performance with GPU acceleration.