Cristallo
Guest
I tried your lisp on the lines.
I designed 810 lines by copying them on 20 times, for a total of 16200 lines
then I launched the lisp that correctly deleted me 810x19=15390 duplicate lines.
All the process, on my car, lasted 59 seconds.
Obviously comparing faces compared to lines will be a more burdensome compute, but on your a test like that it takes?
I bet an idea, maybe you can reduce the number of comparisons. . .
if when you start with the first idx you make him make a selection at crossing that has as window the first +totx/toty and -totx/toty (I try to explain better, a window of the size you prefer, whose median point are the coordinates of the summit of the idx1)?
In this way, within your total selection, you will have a partial selection of entities passing or approaching the first entity, which will not contain all the others at all. at this point the comparison you do only on a limited number of entities, not on all xmila that maybe are far away and have nothing in common with idx1.
I repeat, it's just an idea, from here to implement it, it passes.
I designed 810 lines by copying them on 20 times, for a total of 16200 lines
then I launched the lisp that correctly deleted me 810x19=15390 duplicate lines.
All the process, on my car, lasted 59 seconds.
Obviously comparing faces compared to lines will be a more burdensome compute, but on your a test like that it takes?
I bet an idea, maybe you can reduce the number of comparisons. . .
if when you start with the first idx you make him make a selection at crossing that has as window the first +totx/toty and -totx/toty (I try to explain better, a window of the size you prefer, whose median point are the coordinates of the summit of the idx1)?
In this way, within your total selection, you will have a partial selection of entities passing or approaching the first entity, which will not contain all the others at all. at this point the comparison you do only on a limited number of entities, not on all xmila that maybe are far away and have nothing in common with idx1.
I repeat, it's just an idea, from here to implement it, it passes.