cacao: Compute And Control for Adaptive Optics

cacao is a real-time computation engine for adaptive optics control. It is written in C for low latency, uses multi-core CPUs and GPUs, and keeps data in shared-memory image streams with low-latency inter-process signaling.

cacao is a plugin of milk. It is built inside milk's source tree and adds AO modules, CLI commands and loop setup scripts on top of milk.

This repository is the Caltech Optical Observatories fork of cacao-org/cacao.

Documentation

The documentation is published at https://caltechopticalobservatories.github.io/cacao/. Its sources are in docs/:

• Installation
• Running a loop from an example
• Modules
• Development

To build the HTML documentation, run tox -e docs.

Installing

Install the prerequisites first (list). Then:

git clone https://github.com/milk-org/milk.git
cd milk
git clone https://github.com/CaltechOpticalObservatories/cacao.git plugins/cacao-src

cmake -S . -B _build -DCMAKE_BUILD_TYPE=Release -DINSTALLMAKEDEFAULT=ON
cmake --build _build -j "$(nproc)"
sudo cmake --install _build

Run cmake from the milk source root as shown. Configuring from inside _build/ (cmake .., or milk's compile.sh) silently leaves cacao out of the build.

Then set MILK_ROOT, MILK_INSTALLDIR, PATH and PKG_CONFIG_PATH, and set up the shared-memory directory, as described in the installation guide.

To build in Docker instead: docker build -t cacao .

Getting started

cacao-loop-deploy -h                      # list example configurations
cacao-loop-deploy -c scexao-vispyr-bin2   # copy an example into the current directory
cacao-loop-deploy -r scexao-vispyr-bin2   # run its setup

See Running a loop from an example.

Help and issues

Report bugs on the issue tracker.