The Multithreaded Adaptive Optics Simulator (MAOS) is an end-to-end adaptive optics (AO) simulator. It can simulate many kinds of astronomical AO systems, including single conjugate AO (SCAO), multi-conjugate AO (MCAO), laser tomography AO (LTAO), multi-object AO (MOAO), and ground layer AO (GLAO). It is written in C, runs on multi-core CPUs, and can optionally use Nvidia GPUs.
User guide
- Introduction – what MAOS models and how it is organized.
- Compile the Code – download, dependencies, configure options, GPU and MATLAB support, the job monitor, and the Python scripts.
- Run simulations – command line options, configuration files, and environment variables.
- Examples – walk-through of the TMT NFIRAOS MCAO and SCAO example configurations.
- Config File List – the complete list of configuration files.
- Simulation Results – the output files and how to read and plot them.
- NFIRAOS Performance – performance of the TMT NFIRAOS baseline.
- Algorithms – notes on DM influence functions, DM hysteresis, and physical optics beam propagation.
- Sky Coverage – the sky coverage post-processor skyc.
Developer guide
Publications
Historical benchmarks
Time per simulation step for the TMT NFIRAOS baseline (30 m aperture, 6 LGS, dual order 60x60 DMs), measured on hardware of the time:
- Dual Intel Xeon W5590 at 3.33 GHz: 1.3 s with 8 threads.
- Intel Core i7-2600 at 3.4 GHz: 1.9 s with 4 threads.
- Intel Core i7-2600 at 3.4 GHz: 1.67 s with 8 hyper-threads (4 physical cores).
- One Nvidia GTX 580 GPU: 0.2 s.
- Eight Nvidia GTX 580 GPUs: 0.03 s.
Source code and author
The source code is on GitHub: lianqiw/maos. This manual is at lianqiw.github.io/maos.
- Author
- Lianqi Wang (LinkedIn) at TMT International Observatory.