Dear all, every time I try to make use of the optical routine of siesta I'm impressed by its crawling. If a single scf cycle takes 1 hour, the optical calculation took 30.9 days! This is the opposite of what happens in plane-wave codes, where the proportion of scf-vs-optical time consumption is the opposite: hours for scf, and minutes for optical EPSIMG.
So I'm asking the developers: is there any technical reason for a such frustrating slowness of siesta-optical? Is there any space for substantially improving this routine? thank you, RG
