Setup and install the SDK

This document serves as a guide for installing the Quantum Rings SDK. You can choose to install the SDK in either CPU-only mode or GPU-enabled mode, depending on your system configuration. Before starting the installation process, please review the sections that outline the minimum system requirements and supported GPU architectures.

Quantum Rings SDK also supports the CUDA-Q platform in both CPU only or GPU enabled modes.

Choose the version that is appropriate for your hardware, task, and programming skills using the following table.

Choosing the correct version

If you are …

No NVIDIA GPU (CPU Only mode)
NVIDIA GPU Available (or)
Using GB10 Grace Blackwell Superchip

If you need Qiskit integration
trained in
or
want to use Qiskit-like API

Installation Procedure for CPU only Mode

Installation Procedure for GPU Enabled Mode
In addition, install the
Quantum Rings toolkit for Qiskit
Installing the toolkit for Qiskit
interested in using
the CUDA-Q Platform

Installing the Quantum Rings SDK for NVIDIA CUDA-Q

Installing the Quantum Rings SDK for NVIDIA CUDA-Q

Not Supported

Note

Install either the CPU version or the GPU version, and not both. When you install the GPU version, support for CPU-based circuit execution is built in. Therefore, the CPU version does not need to be installed separately when installing the GPU version.

Minimum System Requirements

A system with a configuration better than the minimum requirements is advised. Lower configurations may affect the number of qubits that can be supported and may perform poorly.

  • Operating systems supported:

    • Windows 11 Pro

    • Google Colab

    • Debian GNU/Linux 12 (bookworm)

    • OpenSUSE Tumbleweed - Version 20240415

    • Oracle Linux 9.3

    • Ubuntu 22.04.4 LTS

    • macOS Sequoia (Only CPU mode and official Python builds are supported. For more information, visit www.python.org.)

    • Windows 11 Pro + WSL2 based Linux Installations

  • 64-bit x86 CPU (14 cores 20 logical processors recommended) on Windows or Linux platforms

  • 64-bit Intel or Apple Silicon CPUs on Apple PCs

  • NVidia GB10 Grace Blackwell Superchip

  • 32 GB Installed physical memory

  • 18 GB Available physical memory, actual memory requirement depends upon the Quantum Circuit.

  • 64-bit Python version 3.11, 3.12, 3.13, or 3.14

NVIDIA GPU Support

  • Amphere, compute capabilities 8.0, 8.6

  • Hopper, compute capability 9.0

  • Blackwell, compute capability 10.0

  • or later

  • CUDA Runtime 12.9 or 13.0 is required on Windows and Linux systems

  • CUDA Runtume 13.0 on GB10 Grace Blackwell Superchip

A minimum of 4 GB global memory on the GPU and a compute capability of 8.0 is required to run the SDK. The actual amount of memory required depends upon the number of qubits used and the gate operations involved.

If you are a Windows user, you can also run the SDK from a WSL based Linux instance. Some of the Qiskit Community projects such as Nature package uses libraries (pyscf) which are better run on Linux. For a tutorial on installing Linux on Windows using WSL, you may refer: WSL Installation Tutorial