ADC-Benchmark

Adaptive Dissipation Control — Comprehensive benchmark suite for superconducting qubits.

Features

• Baseline, Dynamical Decoupling (XY8), ADC, Hybrid
• Temperature-dependent optimization
• Bootstrap confidence intervals
• Quick/Full modes for CI vs. full experiments

Requirements

• Python 3.10+
• Recommended: conda (qutip is easiest to install from conda-forge)

Quick install (recommended)

# create env (recommended)
conda create -n adcbench -c conda-forge python=3.10 qutip matplotlib numpy
conda activate adcbench

Or with pip (may fail or be slow for qutip):

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Run

• Quick smoke run (fast, runs 50 mK quick mode, writes adc_temperature_sweep.png):

python -m src --quick

Or run the script directly:

python src/adc_benchmark.py --quick

• Full benchmark:

python -m src

Or:

python src/adc_benchmark.py

CLI flags

Flag	Description
--quick	Run abbreviated benchmark (50 mK only, ~1-2 min)
--output PATH	Output path for the plot (default: adc_temperature_sweep.png)
--save-results	Save results as both JSON and CSV to --results-dir
--results-dir DIR	Directory for saved results (default: results)
--json	Save results as JSON only
--csv	Save results as CSV only

Example — save JSON after a quick run:

python -m src --quick --json

Testing & CI

• A lightweight smoke test is provided in tests/. Run it locally with:

pytest -q tests/test_smoke.py

• GitHub Actions runs the smoke test in a small conda environment.

License

• MIT (see LICENSE file)