Qplacer: Deep learning toolkit-enable Superconducting Physical Qubit Placement

What it is doing

An efficient tool leveraging deep learning techniques to optimize nonlinear superconducting physical qubit layouts. Developed based on Dreamplace with flexible deep learning toolkits, it runs on both CPUs and GPUs for versatile deployment.

Who will benefit

Researchers on quantum hardware device design, quantum device manufacture, and quantum design automation.

Reference Flow

Sample Result

Publications

• Junyao Zhang, Hanrui Wang, Qi Ding, Jiaqi Gu, Reouven Assouly, William D Oliver, Song Han, Kenneth R Brown, Hai Li, Yiran Chen, "Qplacer: Frequency-Aware Component Placement for Superconducting Quantum Computers", arxiv, 2023 (preprint)

Installation

This project is best built using a Docker container to streamline the process and manage all dependencies efficiently. CMake is adopted as the makefile system.

1. Docker Installation: Begin by installing Docker. Choose the appropriate version for your operating system:

   • Windows
   • Mac
   • Linux

2. Enabling GPU Features (Optional): For GPU support, install NVIDIA-docker. This step can be skipped if GPU features are not needed.

3. Clone the Repository: The repository can be cloned in one of two ways:

   • Full Recursive Clone: This option clones the repository and all its submodules:

     git clone --recursive https://github.com/JJJayyyy/Qplacer.git
     cd Qplacer

   • Selective Submodule Clone: This option only clones the necessary submodule of the project, which can save time and disk space:

     git clone https://github.com/JJJayyyy/Qplacer.git
     cd Qplacer
     
     git submodule init thirdparty/cub
     git submodule init thirdparty/pybind11
     git submodule init thirdparty/Limbo_qplacer_backup/limbo/thirdparty/OpenBLAS
     git submodule update

4. Build the Docker Image: Use the following command to build the Docker image, you can replace jz420 with your name:

   docker build . --file Dockerfile --tag jz420/qplacer:cuda

5. Launching the Container: Depending on your setup (with or without GPU), use following command to start the Docker container and enter its bash environment.

   • With GPU on Linux:

     docker run --gpus 1 --name qplacer_container -it -v $(pwd):/Qplacer -w /Qplacer jz420/qplacer:cuda bash

     remove option --gpus 1 to disable GPU.

6. Setup Environment: To set up the necessary environment and install the required packages:

   Activate the qplacer_env environment:

   conda activate qplacer_env

   Install all the dependencies listed in the requirements.txt file:

   pip install -r requirements.txt

7. Build: Navigate to the qplacer directory to start the build process. Execute the compile.sh script:

   ./compile.sh
   

   Where build is the directory where to compile the code, and qplacer/operators is the directory where to install operators.

   • Conditional Build Option: If all submodules have been cloned, there is a option to use the latest operators from DreamPlace by executing the following command:

     ./compile.sh OFF

   Warning: Using the latest operators may lead to compatibility issues if there are changes that are not compatible with the current state of Qplacer.

   build directory can be removed after installation if you do not need incremental compilation later. To clean, go to the root directory.

   rm -r build
   

Get Benchmarks

To get quantum topology benchmarks, please refer the qplacer/qplacer_example.ipynb. The benchmark files and placement configuration files will be saved in directory qplacer/benchmark and qplacer/test, respectively

Run Placer

Before running, make sure the benchmarks have been created and the python dependency packages/operators have been installed/built successfully.

Navigate to directory qplacer and run qplacer_engine/Placer.py with json configuration file for full placement.

python qplacer_engine/Placer.py test/grid-25/wp_wf/grid-25_wp_wf.json

Contact

Junyao Zhang Email, Github issue

Citation

@misc{zhang2024qplacer,
      title={Qplacer: Frequency-Aware Component Placement for Superconducting Quantum Computers}, 
      author={Junyao Zhang and Hanrui Wang and Qi Ding and Jiaqi Gu and Reouven Assouly and William D. Oliver and Song Han and Kenneth R. Brown and Hai "Helen" Li and Yiran Chen},
      year={2024},
      eprint={2401.17450},
      archivePrefix={arXiv},
      primaryClass={quant-ph}
}