I realized that as example 58% swing on 1/8 is not on the same grid as 58% swing on 1/16 grid. why?????

1/16th is half the length of 1/8. So the Swing should change accordingly.
58% on 1/8 would be simular as 64% on 1/16th
Not sure it really works like that... Maybe you want swing only on the alternating notes, thus set on the 1/8ths

In a typical groove with swung sixteenth notes, the eighth notes are straight (equal length). It's the sixteenth notes that fall in between the eighth notes that are held back a little bit. So regardless of numbers, this is a different grid than any kind of eighth note swing.

In an eighth note swing, the eighth notes would be your pulse and there probably wouldn't be many (or any) sixteenth notes. When there are sixteenth note flourishes in this kind of groove, they don't necessarily line up with the "grid", which means software quantize might get it wrong.

IE: if the denominator of the time, as nystul says 'the pulse' is 8th notes they would be straight, of course. and the quantize swing value is going to apply to 16ths. just as in eg., 4/4 time it makes no sense to have swung quarters.

if the denominator is quarter notes I don't know why you would have differing swing grids of 8ths v. 16ths. there is such a thing as swung 16ths where the pulse is quarter, in addition to swinging at the 8th note level, etc. I have never relied on a machine to do it; I don't know why, mechanically speaking a program would automatically be inconsistent if this is the case.