LISA EMRI/IMRI Figures of Merit

Overview

This repository provides comprehensive tools for computing Figures of Merit (FoMs) for Extreme Mass Ratio Inspirals (EMRIs) and Intermediate Mass Ratio Inspirals (IMRIs) observable by the Laser Interferometer Space Antenna (LISA) mission. If you use the results or codes contained here please cite our work also available here.

What are EMRIs and IMRIs?

EMRIs occur when stellar-mass compact objects (1-100 M☉) spiral into supermassive black holes (10⁴-10⁷ M☉), providing unique laboratories for studying the galactic centers and testing general relativity in the strong-field regime. IMRIs involve intermediate-mass black holes and offer insights into black hole formation and growth across cosmic time.

What This Repository Does

We quantify LISA's capability to:

• Detect EMRIs and IMRIs across different mass ranges and redshifts
• Characterize source parameters (masses, spins, eccentricity, distance) with precision
• Assess impact of detector sensitivity degradation and mission duration on science objectives

Interactive Exploration

Try it yourself! Explore how LISA sensitivity changes affect EMRI/IMRI detection and parameter estimation using the interactive notebook or the interactive website. No installation required - just click to launch in your browser.

Note: Binder environments are pre-built using GitHub Actions, so launch time is typically under 1 minute. If you see a longer build time, the container may be updating with recent changes.

Using This Work

If you use any material from this repository (code, data, figures, or results) in your research, please cite our work. See the Citation section below for details.

---

Key Results

Our analysis provides quantitative insights into LISA's EMRI/IMRI science capabilities across the following parameter space:

Parameter Space Coverage

EMRI and IMRI parameter space: primary vs secondary mass configurations

Detection Horizon

Maximum detection redshift as a function of system parameters

Parameter Estimation Precision

Measurement precision across the mass parameter space

---

Reproducing this work

To reproduce the full analysis pipeline on your own GPU-enabled system, follow the installation instructions below. The pipeline uses GPU acceleration for efficient waveform generation and Fisher matrix calculations.

Pre-built Container: A ready-to-use Singularity container is available here. We also provide Singularity container instructions to create your own container below.

Quick Start with Pre-built Container

Download and use the ready-made container here:

# Download container (if needed)
# Available at: https://public.spider.surfsara.nl/project/lisa_nlddpc/emri_fom_container/

# Test the container
singularity exec --nv fom_final.sif python -m unittest test_waveform_and_response.py

# Run pipeline analysis
cd pipeline
singularity exec --nv ../fom_final.sif python pipeline.py \
    --M 1e6 --mu 1e1 --a 0.5 --e_f 0.1 --T 4.0 --z 0.5 \
    --psd_file TDI2_AE_psd.npy --dt 10.0 --use_gpu \
    --N_montecarlo 1 --device 0 --power_law --repo test_acc \
    --calculate_fisher 1

Building Your Own Container

Request GPU Node

Example for Spider HPC:

# Request GPU partition
srun -p gpu_a100_22c --mem 64G -G a100:1 -c 2 --pty bash

Build Container

Create a container from the definition file:

# Build final container
singularity build --nv --fakeroot fom_final.sif fom.def

# OR build editable sandbox for development
singularity build --sandbox --nv --fakeroot fom fom.def

Customize Editable Container

Open a shell in the editable container:

singularity shell --writable --nv --fakeroot fom

Install additional packages inside the container:

# Update pip
python -m pip install --upgrade pip

# Install core dependencies
python -m pip install --no-cache-dir nvidia-cuda-runtime-cu12 astropy \
    eryn fastemriwaveforms-cuda12x multiprocess optax matplotlib scipy \
    jupyter interpax numba Cython lisaanalysistools tabulate scienceplots \
    healpy pandas filelock

# Test FEW installation
python -c "import few; few.get_backend('cuda12x'); print('FEW installation successful')"

# Set compilers for building GPU packages
unset CC CXX CUDACXX
export CC=/usr/bin/gcc
export CXX=/usr/bin/g++
export CUDACXX=/usr/local/cuda/bin/nvcc
export NVCC_PREPEND_FLAGS='-ccbin /usr/bin/g++'

# Clone and install EMRI-FoM packages
git clone https://github.com/lorenzsp/EMRI-FoM.git emri_fom_temp
cd emri_fom_temp/lisa-on-gpu/
python setup.py install
cd ../StableEMRIFisher-package/
python -m pip install .
cd ..

# Run tests
python -m unittest test_waveform_and_response.py

Convert sandbox to final image:

singularity build fom_final.sif fom

---

Installation with Conda

Prerequisites

• CUDA-capable GPU with compute capability ≥ 7.0
• CUDA Toolkit 12.x with nvcc compiler
• Anaconda or Miniconda
• Linux or macOS (Windows not currently supported due to compiler limitations)

Step 1: Create Environment and Install FEW

Fast EMRI Waveforms (FEW) provides GPU-accelerated waveform generation:

# Create environment with FEW
conda create -n fom_env -c conda-forge -y --override-channels \
    python=3.12 fastemriwaveforms-cuda12x

# Activate environment
conda activate fom_env

# Install additional dependencies
pip install tabulate markdown pypandoc scikit-learn healpy \
    lisaanalysistools seaborn corner scipy tqdm jupyter \
    ipython h5py requests matplotlib eryn Cython

Test FEW installation:

import few
few.get_backend("cuda12x")  # Should complete without errors

Step 2: Install Fisher Information Package

cd StableEMRIFisher-package/
pip install .
cd ..

Step 3: Install LISA Response (lisa-on-gpu)

First, locate your CUDA compiler and add it to PATH:

# Find nvcc location (typically in /usr/local/cuda-*/bin/)
export PATH=$PATH:/usr/local/cuda-12.5/bin/

Then install the response package:

cd lisa-on-gpu
python setup.py install
cd ..

Verify installation:

from fastlisaresponse import ResponseWrapper  # Should import without errors

Step 4: Test Installation

Run the test suite:

# Test waveform and response
python -m unittest test_waveform_and_response.py

# Test pipeline with minimal configuration
cd pipeline
python pipeline.py --M 1e6 --mu 1e1 --a 0.5 --e_f 0.1 --T 4.0 --z 0.5 \
    --psd_file TDI2_AE_psd.npy --dt 10.0 --use_gpu --N_montecarlo 1 \
    --device 0 --power_law --repo test_acc --calculate_fisher 1

If all tests pass, your installation is complete! ✓

---

Alternative: Python Virtual Environment

For systems without Singularity or conda, use a standard Python virtual environment:

# Request compute node (HPC example)
srun --partition=short --time=12:00:00 --pty bash -i -l

# Create and activate virtual environment
python -m venv fom_venv/
source fom_venv/bin/activate

# Install packages
python -m pip install --upgrade pip
python -m pip install --no-cache-dir nvidia-cuda-runtime-cu12 astropy \
    eryn fastemriwaveforms-cuda12x multiprocess optax matplotlib scipy \
    jupyter interpax numba Cython lisaanalysistools tabulate scienceplots \
    healpy pandas filelock

# Launch Jupyter (for remote work)
jupyter lab --ip="*" --no-browser

# On local machine, create SSH tunnel:
# ssh -NL 8888:wn-la-01:8888 spider

---

Binder Performance Optimization

This repository uses pre-built Docker containers to significantly reduce Binder launch time:

• GitHub Actions automatically builds and pushes a Docker image whenever requirements.txt or key files change
• Binder pulls the pre-built image from GitHub Container Registry instead of building from scratch
• Result: Launch time reduced from 5-10 minutes to typically under 1 minute

The workflow is defined in .github/workflows/binder-build.yml and uses repo2docker to create the same environment that Binder would build.

For Maintainers

After updating requirements.txt:

1. Commit and push changes to the main branch
2. GitHub Actions will automatically rebuild and push the Docker image
3. Subsequent Binder launches will use the updated pre-built image

To manually trigger a rebuild, go to the Actions tab and click "Run workflow".

---

Contributing

We welcome contributions! Please feel free to submit issues, feature requests, or pull requests.

Citation

If you use the results or code provided in this repository, please cite:

@article{Speri:2026ade,
    author = "Speri, Lorenzo and Duque, Francisco and Barsanti, Susanna and Santini, Alessandro and Kejriwal, Shubham and Burke, Ollie and Chapman-Bird, Christian E. A.",
    title = "{Quantifying the Scientific Potential of Intermediate and Extreme Mass Ratio Inspirals with the Laser Interferometer Space Antenna}",
    eprint = "2603.17072",
    archivePrefix = "arXiv",
    primaryClass = "astro-ph.IM",
    month = "3",
    year = "2026"
}

If you use the electromagnetic observations presented in this work please read the paper, in particular Appendix A and also cite the corresponding papers.

License

This project is licensed under the Apache License - see the LICENSE file for details.

Contact

For questions or support, please open an issue on GitHub or contact the maintainers.

---