arXiv:2507.18379

From Ising to Potts: Physics-inspired Potts machines of coupled oscillators for low-energy sampling and combinatorial optimization

Yi Cheng, Zongli Lin

Published Jul 23, 2025• 0 citations

Abstract

Repository containing the code to reproduce the experiments in the paper "From Ising to Potts: physics-inspired Potts machines in coupled oscillators for low-energy sampling and combinatorial optimization". It contains four standalone Python scripts: Max_2_cut.py — OPM solver for Max-2-Cut Max_3_cut.py — OPM solver for Max-3-Cut Max_4_cut.py — OPM solver for Max-4-Cut Max_5_cut.py — OPM solver for Max-5-Cut The benchmark graph instances G1–G5 are included as text files (G1.txt–G5.txt) in the same folder. Requirements Python 3.9+ and the packages listed in requirements.txt (notably: numpy, scipy, matplotlib, torch, torchsde). How to run (Jupyter / terminal) After installing dependencies: In a Jupyter notebook: %run Max_2_cut.py %run Max_3_cut.py %run Max_4_cut.py %run Max_5_cut.py Or in a terminal: python Max_2_cut.py (and similarly for Max-3/4/5) The scripts generate the output figures in the working directory (e.g., box plots summarizing solution quality across instances). GPU recommendation For best performance, we recommend running the code on a GPU. The runtime numbers reported below were measured on NVIDIA GeForce RTX 2080 Ti (CUDA 12.8; driver 570.124.06). Actual wall-clock time may vary with hardware and system load. Reproducibility and runtime For a faster sanity check, start with seed_num = 1 (typical wall-clock times to finish all instances G1–G5: Max-3-Cut ~12 min, Max-4-Cut ~16 min, Max-5-Cut ~24 min).To reproduce the results reported in the paper, use seed_num = 20.

DOI: 10.48550/arxiv.2507.18379

Related papers

Potts model solver with coherent Ising machines

Takahiro Inagaki, Kensuke Inaba, Koji Igarashi +9 more

The Japan Society of Applied PhysicsJul 10, 20190 citations

Kernpiler: Compiler Optimization for Quantum Hamiltonian Simulation with Partial Trotterization

Ethan Decker, Lucas Goetz, Evan McKinney +9 more

Description This artifact contains the core implementation of Kernpiler, a compiler framework for optimizing quantum circuits, and supports full reproducibility of all experimental results presented in the associated paper. The artifact includes all code, data pipelines, and scripts required to regenerate Figures 5–11. System Used for Data Collection NVIDIA A100 GPU with 80GB memory AMD EPYC 9654P 96-core processor x86_64 Linux system Python 3.13.5 Experiments may be computationally intensive, but they are fully parallelizable across multiple devices. Installation Clone or download the repository and navigate to the project directory. Create a virtual environment: python3 -m venv validate source validate/bin/activate Install dependencies: python -m pip install -r requirements.txt Install torch-scatter: python -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-2.10.0+cu128.html Experiment Workflow All experiment scripts are located in: src/compiler/optimization_passes/experiments Each figure can be reproduced by running its corresponding data collection and graphing scripts: Figure 5exp_gatecount_datacollection.py→ graph using:graph_data_scripts/graph_absolute.py Figure 6exp_partition_scaling_datacollection_o1exp_partition_scaling_datacollection→ graph using:graph_data_scripts/graph_o1_o2_side_by_side.py Figure 7exp_runtime_per_pass.py→ graph using:graph_data_scripts/graph_runtime_per_pass.py Figure 8exp_partition_scaling_datacollectiono1_phoenixFT→ graph using:graph_data_scripts/graph_firstorder_scalingFT.py Figure 9exp_partitionalgvsrandom.py→ graph using:graph_data_scripts/graph_partition_vs_random.py Figure 10exp_scaling_data_rewriteradius.py→ graph using:graph_data_scripts/graph_scalingdata.py Figure 11exp_error_scaling_systemsize.py→ output generated directly (no additional graph script required) Execution Notes All experiments are independent Parallel execution is supported Runtime varies depending on system size and hardware

arXiv:2504.07214Proceedings - International Symposium on Computer Architecture/ProceedingsJun 1, 20261 citation

Sensing with discrete time crystals

Moon, Leo, Schindler, Paul, Smith, Ryan +4 more

# Sensing with discrete time crystals Data Repository This repository contains experimental data, simulation results, and analysis code for Discrete Time Crystal (DTC) sensing research. The data is organized by figure numbers corresponding to the main manuscript and supplementary materials. ## Repository Structure ``` DTC_sensing_Zenodo_Repo/ │ ├── data/ # All experimental and processed data │ ├── raw/ # Raw experimental data files │ │ ├── fig-1/ # Figure 1 raw data │ │ ├── fig-2/ # Figure 2 raw data (fig-2b, fig-2c) │ │ ├── fig-3/ # Figure 3 raw data (fig-3a, fig-3b, fig-3c) │ │ ├── fig-4/ # Figure 4 raw data │ │ └── fig-S1/ through fig-S18/ # Supplementary figures raw data │ │ │ └── processed/ # Processed and analyzed data │ ├── fig-3/ # Processed data for Figure 3 │ └── fig-S14/ # Processed data for Figure S14 │ ├── src_code/ # Source code for analysis and simulations │ ├── simulations/ # Simulation code │ │ ├── SingleToneDTC/ # Single-tone DTC simulations │ │ ├── TwoToneDTC/ # Two-tone DTC simulations │ │ ├── helper.py # Helper functions │ │ ├── setup_system.py # System setup utilities │ │ └── systems.yaml # System configuration parameters │ │ │ └── utils/ # Utility functions │ ├── plotting_utils.py # Plotting and visualization utilities │ └── simulation_utils.py # Simulation utilities │ └── visual_elements/ # Code and notebooks for figure generation ├── main/ # Main manuscript figures │ ├── fig-1/ # Figure 1 generation code │ ├── fig-2/ # Figure 2 generation code │ ├── fig-3/ # Figure 3 generation code │ └── fig-4/ # Figure 4 generation code │ └── supmat/ # Supplementary material figures └── fig-S1/ through fig-S18/ # Supplementary figure generation code ``` ## Data Description ### Raw Data (`data/raw/`) The raw data directory contains experimental measurements organized by figure number: - **File Formats:** - `.mat` files: MATLAB data files containing experimental measurements from the Proteus P9484M NMR spectrometer - `.csv` files: CSV files containing contour and magnetometer data (fig-S1) - `.hdf5` files: HDF5 files containing simulation results (fig-S13-S16) - `.npy` files: NumPy array files - **Data Organization:** - **Main Figures (fig-1 through fig-4):** Core experimental results - **Supplementary Figures (fig-S1 through fig-S18):** Additional experimental data, simulations, and analysis ### Processed Data (`data/processed/`) Processed data contains analyzed and pre-computed results: - **File Formats:** - `.npy` files: NumPy arrays containing processed experimental data - **Available Processed Data:** - `fig-3/`: Processed data for Figure 3 (fig-3a, fig-3b) - `fig-S14/`: Processed data for Figure S14 (AC phase scans, frequency scans, Vpp scans) ## Source Code ### Simulation Code (`src_code/simulations/`) - **SingleToneDTC/**: Code for single-tone DTC simulations - **TwoToneDTC/**: Code for two-tone DTC simulations - **systems.yaml**: Configuration file defining system parameters (N, rmin, rmax, gamma, theta, tau, etc.) - **helper.py**: Helper functions for simulations - **setup_system.py**: System setup and initialization utilities ### Utility Functions (`src_code/utils/`) - **plotting_utils.py**: Functions for data visualization, including matplotlib configuration and plotting routines - **simulation_utils.py**: Utilities for running and analyzing simulations ## Figure Generation ### Visualization Code (`visual_elements/`) Each figure directory contains: - **Jupyter notebooks** (`.ipynb`): Interactive notebooks for data analysis and figure generation OR - **Python scripts** (`.py`): Standalone scripts for figure generation - **Output files** (`.pdf`, `.svg`): Generated figure files ## Requirements All required Python packages are listed in `requirements.txt`. Key dependencies include: - **NumPy** (2.3.5): Numerical computations - **SciPy** (1.16.3): Scientific computing and data analysis - **Matplotlib** (3.10.7): Plotting and visualization - **h5py** (3.15.1): HDF5 file handling - **Pandas** (2.3.3): Data manipulation - **Jupyter/IPython**: Interactive notebooks - **PyYAML** (6.0.3): Configuration file parsing To install dependencies: ```bash pip install -r requirements.txt ``` ## Data Access ### Reading Raw Data Raw experimental data is stored in MATLAB `.mat` format. To read these files in Python: ```python from scipy.io import loadmat data_dict = loadmat('path/to/file.mat') ``` ## Usage 1. **Install dependencies**: `pip install -r requirements.txt` 2. **Navigate to figure directory**: Each figure has its own directory in `visual_elements/` 3. **Run analysis**: Open the Jupyter notebook (`.ipynb`) or Python script (`.py`) for the desired figure 4. **Data paths**: The code assumes data is in the `data/` directory relative to the repository root ## Notes - Raw data files are named with timestamps (e.g., `2024-06-06-125713_Proteus.mat`) indicating when measurements were taken - Some figures contain subdirectories (e.g., `fig-2b/`, `fig-2c/`) for different panels - The repository structure follows the organization of figures in the associated publication

Nov 28, 20251 citation

An FPGA Implementation of Fast 2D-Ising-Model Solver for the Max-Cut Problem

英範業天, 正之廣本, 高史佐藤

IEICE technical report. SpeechFeb 29, 20162 citations

An FPGA Implementation of Fast 2D-Ising-Model Solver for the Max-Cut Problem

Hidenori Gyoten, Masayuki Hiromoto, Satō Takashi

IEICE Technical Report; IEICE Tech. Rep.Feb 22, 20162 citations

Code and datasets for "Shortcuts to adiabaticity with a quantum control field"

Emma C. King, Giovanna Morigi, Raphaël Menu

Shortcuts to adiabaticity with a quantum control field — Code & data========================================================== This repository contains a Jupyter notebook and supporting data files (csv) to reproduce the numerical results and figures in our manuscript (arXiv 2508.06304), which investigates adiabatic shortcuts mediated by a quantum control field.The code implements a Landau–Zener two-level system coupled to a spectator qubit, time evolution of the full system (LZ qubit + spectator) via QuTiP, parameter sweeps (parallelized with joblib), and plotting utilities for the manuscript figures. Related publication----------------------------This code is complementary to the preprint: https://arxiv.org/abs/2508.06304. Requirements---------------------The simulations were developed for and tested on Python 3.9+. Recommended packages and minimum versions: - Python >= 3.9- numpy >= 1.22- scipy >= 1.8- matplotlib >= 3.5- qutip >= 4.6 (install a recent stable qutip; some installations require `pip install qutip` or `conda` channels)- joblib >= 1.2- tqdm >= 4.60- pandas (optional, for CSV handling)- jupyterlab / notebook (for interactive use) Running the notebook (interactive) Open the notebook with JupyterLab or Jupyter Notebook. Execute cells sequentially in the order indicated by the notebook outline.The sections “Definitions” and “Figure imports and setup” must be executed before generating any data or figures, as they define all operators, and helper functions. The notebook is organized into separate sections for each figure of the manuscript, allowing you to re-run only the figure(s) of interest. Note: Figure 4 (Density plot) is configured to load precomputed data files to reduce computational cost. To reproduce Figure 4 without recomputing the full dataset, ensure the required .csv files are downloaded and located in the same directory as the notebook. To explore different parameter regimes: Modify the relevant parameter definitions (e.g., anneal_window, g, epsilon, omega_c, x0, x0_vals, q_vals). To include dissipation of the spectator (ancilla) qubit, set the parameter kappa> 0. For questions about reproducing the figures or about parameter choices used in the paper, contact the corresponding authors listed in the preprint. If you reuse this data/code, please cite the associated preprint and this code archive.

Feb 22, 20260 citations