Re: Orientation of rotationally invariant bit vectors
- From: Nicholas King <zeddie@xxxxxxxxxxxxxxxx>
- Date: Thu, 18 Oct 2007 01:18:09 +1000
acd wrote:
Here are some numbers:I would be inclined to only store one rotation and only do one comparison
when synthesizing to Cyclone II the 2*minimum-tree + comparison
requires:
f
n locic cells delay
8 168 18.4
10 270 24.3
12 396 25
I think there should be a far better way.
Andreas
(I'd be inclined to store one rotation and impose an ordering criteria) then you can rotate an input pattern and order the rotations by this criteria and choose the 'largest'.
One criteria I would suggest is using number of consecutative 1's or O's with having 1^n or O^n as a higher priority than the other (whichever one you choose is arbitary.
It's just a late night thought.
Good luck with the problem.
Nicholas King.
.
- References:
- Prev by Date: Re: Orientation of rotationally invariant bit vectors
- Next by Date: Re: Would like to find study group for computer science comprehensive exams.
- Previous by thread: Re: Orientation of rotationally invariant bit vectors
- Next by thread: Re: Would like to find study group for computer science comprehensive exams.
- Index(es):
Relevant Pages
|