search

rodluger / Limbdark.jl

Analytical transit light curves for limb darkened stars
MIT License
14 stars 4 forks source link

Increase lmax to 15 in Figure 13 #55

Closed rodluger closed 6 years ago

rodluger commented 6 years ago

I recommend increasing the order of the nonlinear fit from 6 to ~15 in Figure 13. This significantly reduces the error while not increasing the run time that much. With l=15, our run time is about 4x faster than batman and we get 7x smaller error. Also, the statement in the caption of Figure 9 isn't quite right, as the accuracy of the fit keeps increasing above l=6 (although it's hard to see on a linear scale; it would be clearer if you plotted it in log space). Somewhere between l=15 and l=20 this turns over, however, since we hit the roundoff error noise floor.

ericagol commented 6 years ago

@rodluger Great! It also looks like you have improved the computational time significantly as well. What allowed for that?