dist_qft

Distributed Quantum Fourier Transform (dQFT)

Installation

Clone this repository.

Installation inscructions are similar to github.com/Quantinuum/pytket-dqc#installing-pytket-dqc#installing-pytket-dqc, but with additional recommendations below to avoid installation issues.

1. Graphviz

apt install libgraphviz-dev pkg-config

2. Python environment

With uv, simply run uv sync from within the dist_qft folder to install python environment.

3. Boost

First, install boost: www.boost.org/doc/libs/1_78_0/tools/build/doc/html/index.html#bbv2#bbv2.installation with suggested prefix: '/usr/local'

Install boost headers: apt install libboost-program-options-dev libboost-dev uuid-dev

4. KaHyPar

Installation instructions are similar to github.com/kahypar/kahypar/blob/1.3.2/README.md#building-kahypar, but with the modifications below for reproducibility and to avoid installation issues.

Use python version 3.11:

sudo add-apt-repository ppa:deadsnakes/ppa
sudo apt update
sudo apt install python3.11 python3.11-dev

Clone KaHyPar version 1.3.2 specifically, using this command:git clone --depth=1 --recursive --branch 1.3.2 https://github.com/kahypar/kahypar.git The KaHyPar directory can go anywhere on your system. The remainder of these installation instructions wil assume it is in /home/ubuntu/.

Fix a bug in the KaHyPar installation:

cd /home/ubuntu/kahypar/python/pybind11
git fetch origin --tags
git checkout v2.12.0

Now run the KaHyPar building instructions as follows:

cd /home/ubuntu/kahypar
mkdir build && cd build
cmake ../ -DCMAKE_BUILD_TYPE=RELEASE -DKAHYPAR_PYTHON_INTERFACE=1 -DPython_EXECUTABLE=/usr/bin/python3.11 -DPYTHON_EXECUTABLE=/usr/bin/python3.11
cd python
make
cp kahypar*.so <path-to-site-packages>

Make sure to use the correct cmake parameters as written above. Make sure to replace <path-to-site-packages> with the correct path for the python site packages used for this repo (dist_qft). For the .venv created by uv, this would be dist_qft/.venv/lib/python3.11/site-packages.

Data and plots

Run main_*_qpus.py to generate data and plot_*_qpus.py to generate graphs from the data.

For reproducibility, run with PYTHONHASHSEED as environment variable. With uv, this can be done with uv run --env-file .env main_few_qpus.py to use the same PYTHONHASHSEED as in the paper.