Installing Qiskit Aer on Ubuntu with Nvidia
Draft notes...
The goal is to install Qiskit Aer 0.17.1 on a 64-bit Linux Ubuntu 24.04.3 LTS with Nvidia GeForce RTX 2080 Super.
Nvidia is already set and running on this machine, Qiskit 2 is also installed but not 2.2 yet.
Aer depends on the main Qiskit package. CUDA (10.1 or newer) and a AER GPU related package are needed to install the GPU supported simulator.
Check if CUDA is installed with nvcc --version
Mine is CUDA 12 and need an upgrade.
If not, follow CUDA Quick Start Guide and CUDA Toolkit 13.1 Downloads . Upgrading CUDA and setting it up on Linux is not always straight forward process. Consult with someone in case of need of help. this guide will not cover this part.
Run the simulators for Cuda to check if it works.
In a Venv install qiskit
python3
Python 3.11.5 (main, Sep 11 2023, 13:54:46) [GCC 11.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import qiskit
>>> print(qiskit.version.get_version_info())
2.2.3
>>>
Install Q Aer
pip install qiskit-aer
Installing collected packages: qiskit-aer
Successfully installed qiskit-aer-0.17.2
and ear-gpu
Successfully installed cudensitymat-cu12-0.3.2 cupauliprop-cu12-0.1.0 cuquantum-cu12-25.11.1 custabilizer-cu12-0.1.0 custatevec-cu12-1.11.0 cutensor-cu12-2.4.1 cutensornet-cu12-2.10.1 nvidia-cublas-cu12-12.9.1.4 nvidia-cuda-runtime-cu12-12.9.79 nvidia-cusolver-cu12-11.7.5.82 nvidia-cusparse-cu12-12.5.10.65 nvidia-nvjitlink-cu12-12.9.86 qiskit-aer-gpu-0.15.1
try out simple program
>>> import qiskit
>>> from qiskit_aer.primitives import SamplerV2