!! This repo will be deprecated !!

We are implementing Q-GPU in a seperate package which is decoupled with qiskit-aer, and will be more portable, stay tuned on qdao

Build from source

To build this project, please follow the instructions in the contribution guidelines. After setting up prerequisite env, run the following commands for compilation.

python ./setup.py bdist_wheel -- -DAER_THRUST_BACKEND=CUDA -DCMAKE_CUDA_COMPILER=${YOUR_NVIDIA_COMPILER_PATH}

To run experiments, please see this repo

To enable multi-GPU simulation, first set up this environment variable

export AER_MULTI_GPU=1

Branches

Branch name	Description
master	w/o compression
compression	w/ compression (num_qubits >= 21)

References

Q-GPU is developped based on QISKit-Aer, 0.7.0. The original version is the baseline of paper.

We use GFC as data compression algorithm in Q-GPU. (With small modifications to adapt to Q-GPU.)