I was facing convergence problems in my 2D system, but after I have used the ALGO = All
tag the convergence problems have gone.
I would like to know how accurate is this option comparing to ALGO = Normal
and ALGO = Fast
options.
1 Answer
ALGO
only refers to the method used to reach convergence, as a result there is no difference in accuracy. ALGO = Normal / Fast
tends to be faster than ALGO = All
, however this can be system / level of theory dependent.
It should be noted, this isn't strictly true since you can optimize to a different ground state minimum. This can result in a different magnetic structure in some cases, but I have never seen this problem in practice.