Table 1: Computational requirements for various orders of adaptive optics systems. D and d are the telescope diameter and the inter-actuator distance given in meters, $N_{\rm DM}$ and $N_{\rm WFS}$ are the number of DMs and WFSs (>1 implies MCAO), n is the total number of actuators, the column " $\vec{c}=E\vec{ s}$'' gives the required real-time processing power in Gflops for the matrix-vector multiplication, the column "comp. E'' gives the required number of floating point operations $\times 10^9$ for computing the reconstructor, and the last column gives the required storage space for E in Gb. A Shack-Hartmann type WFS with a Fried geometry of actuators was adopted for the calculations, so that the total number of measurements is $m\approx 2n$, and the wavefront reconstruction rate was set to 1 kHz. An inter-actuator distance of one meter corresponds roughly to r0 at 2.2 $\mu $m, and d=0.25 approximately to r0 at 0.7 $\mu $m.
D (m) d (m) $N_{\rm DM}$ $N_{\rm WFS}$ n $\vec{c}=E\vec{ s}$ comp. E storage (Gb)
50 1 1 1 2000 8 8 0.03
50 1 2 5 4000 77 61 0.3
100 1 1 1 8000 123 484 0.5
100 1 2 5 16 000 1234 3876 5
50 0.25 1 1 32 000 1974 31 006 8
50 0.25 2 5 63 000 19 739 248 050 80
100 0.25 1 1 126 000 315 83 $2\times 10^6$ 126
100 0.25 2 5 252 000 315 827 $16\times 10^6$ 1263


Source LaTeX | All tables | In the text