Re: about H field output at k point

On Jun 15, 2011, at 5:03 AM, [email protected] wrote:
> Thank you very much for your reply.
>
> I still dont know why i cant get the output if i insist on using the
> output command:
> "(run-te (output-at-kpoint (vector3 0.1 0 0)
>                          fix-hfield-phase output-hfield-z))"
> but i found a solution for the problem by using (list-ref k-points 1)
> instead of (vector3 0.1 0 0), namly:
> "(run-te (output-at-kpoint (list-ref k-points 1)
>                          fix-hfield-phase output-hfield-z))"
> it confused me very much.

Oh, I know what happened.

The problem is rounding errors.  When you create the k-point list with  
"interpolate", the point (0.1,0,0) is generated with a bit of rounding  
error.  When you enter (0.1,0,0) manually there is also a rounding  
(since 0.1 is not exactly representable in binary floating point), but  
it is a different one, so "0.1" in the two cases differs slightly.

On my computer, it worked because I have updated my version of MPB to  
use "vector3-close?" in output-at-kpoint instead of "vector3=", so my  
version allows the vectors to differ slightly by rounding errors.

I should probably try to release a new version of MPB this summer; a  
bunch of little improvements have piled up since the last release.

Steven