Architecture

Simulation Flow

• main() is the entrance into the program. It calls
  • setup_parms(): to setup the parameters of types parms_t and check for errors.
    • setup_parms_gpu(): to setup the GPU usage for various tasks
  • maos_setup(): to setup the AO system geometry and data structs. It calls
    • setup_aper(): to setup the aperture grid and amplitude map (of type aper_t)
    • setup_powfs_init(): to setup the WFS subapertures geometry
    • setup_recon_prep(): to setup reconstruction grid (ploc, aloc, floc, GP, GX, etc.)
    • setup_surf(): to setup NCPA
    • setup_shwfs_phy(): to setup physical optics (dtf, etf, i0, mtch, cog)
    • setup_powfs_neasim(): to setup grad noise model if not using photon noise.
    • setup_powfs_calib(): NCPA calibration
    • setup_recon_GA(): Actuator to gradient
    • setup_recon_GF(): Focus mode to gradient
    • setup_recon_GR(): Radial Zernike modes to gradient
    • ngsmod_prep(): AHST ngsmod geometry
    • setup_recon_saneai(): to setup noise covariance
    • setup_recon_dither_dm(): Common mode dithering setup
    • setup_recon_control(): Setup controller
    • setup_recon_moao(): to setup MOAO
    • setup_powfs_fit(): to setup fitting to WFS subaperture algorithm
    • setup_recon_mvm(): to calculate MVM control matrix
    • gpu_setup_recon_mvm(): calculate MVM control matrix using GPUs
    • gpu_wfsgrad_init(): to setup WFS in GPU.
    • gpu_setup_recon(): to setup recon in GPU.
    • gpu_perfevl_init(): to initialize performance evaluation parameters in GPU.
    • gpu_wfssurf2gpu(): to copy NCPA OPDs to GPU.
    • setup_recon_misc(): to compute PSD, etc.
    • plot_setup(): to plot the geometry (ploc, floc, aloc, amplitude map, etc.).
  • maos_sim(): to start the simulation. It then calls
    • maos_iseed(): to initialize runtime data for each seed including turbulence
    • In event driven simulation mode (PARALLEL=2). It calls the following in parallel each in a loop
      • perfevl() & print_progress(): performance simulation
      • wfsgrad() and shift_grad(): compute gradients and shift to gradlast
      • reconstruct() & filter_dm: wfs reconstruction and servo filtering
    • In parallel simulation, calls maos_isim() for each step. It then calls
      • sim_update_etf(): update sodium profile if needed
      • genatm(): update turbulence if not using frozen-flow
      • In parallel mode (PARALLEL=1)
        • perfevl_pre(): wfsgrad_prep, reconstruct() in parallel and then wait
        • perfevl(): wfsgrad(), in parallel and then wait
        • shift_grad(): copy from grad to gradlast for reconstruct()
        • filter_dm(): servo filtering (dmreal)
      • In serial mode (PARALLEL=0), calls the following in sequence
        • perfevl(): (in closed loop) to evaluate the performance in science field
        • wfsgrad(): to compute WFS gradients
        • reconstruct():
        • shift_grad():
        • filter_dm():
        • perfevl():
      • setup_recon_update_cn2(): Update tomography parameter when needed
      • setup_recon_control(): Update recon parameters when needed.
      • print_progress(): show progress
    • free_simu(): to delete run time structs and close files