EPOC: A Novel Pulse Generation Framework Incorporating Advanced Synthesis Techniques for Quantum Circuits

EPOC is a continuation of the previous work AccQOC. It is a novel pulse generation framework that incorporates advanced synthesis techniques for quantum circuits. On average, EPOC is able to achieve a 31.74% reduction in circuit latency compared with the state-of-the-art pulse generation framework, PAQOC.

The workflow of EPOC

File Preparation

1. Benchmark

   It's recommended to fork the benchmark repository to build the pulse library. The final benchmarks plotted in the paper which are copied from PAQOC are stored in "bench" folder.

2. Pulse Library

   The pulse library is a CSV file that contains the pulse information, which should have following columns:

   • unitary
   • total_time
   • fidelity
   • compilation_time

   A sample pulse library is provided in "pulse_lib.csv".

3. Performance Benchmark

   The performance benchmark is a CSV file that contains the performance of EPOC, which should have following columns:

   • path
   • latency
   • latency_group
   • compilation_time
   • compilation_time_group
   • fidelity
   • fidelity_group

   A sample performance benchmark is provided in "performance_bench.csv".

Environment Setup

1. Create virtual environment

   conda create -n <env_name> python=3.8
   conda activate <env_name>

2. Install quantum optimal control package

   cd PY3_quantum-optimal-control
   python setup.py install

3. Install other dependencies

   cd ..
   pip install -r requirements.txt

Running EPOC

Several parameters are required to run EPOC:

• --maxsize: the maximum partition size
• --dir: the directory of the benchmark
• --filepath: the path to the pulse library
• --dest: the destination to write the performance benchmark

 python run_epoc_bench.py --maxsize=4 --dir="bench" --filepath="pulse_lib.csv"  --dest="performance_bench.csv"           

Code Contributors

Yuchen Zhu, Jinglei Cheng

Contact

Jinglei Cheng [email protected] Yuchen Zhu [email protected]

Citation

@misc{cheng2024epoc,
      title={EPOC: A Novel Pulse Generation Framework Incorporating Advanced Synthesis Techniques for Quantum Circuits}, 
      author={Jinglei Cheng and Yuchen Zhu and Yidong Zhou and Hang Ren and Zhixin Song and Zhiding Liang},
      year={2024},
      eprint={2405.03804},
      archivePrefix={arXiv},
      primaryClass={quant-ph}
}

Acknowledgement

Our implementation is based on BQSKit, PyZx, and GRAPE-Tensorflow.