Re: MPB vs MPBI discrepancy

On Wed, 7 Mar 2007, MHerrera wrote:
> When I run the following ctl file using MPBI vs MPB, I get different results
> between the two, though I believe I should be getting the same output:

I can't reproduce your problem.  I modified your .ctl file to do only 4 
bands, and only at k=0, with a resolution of 10, just to run more quickly, 
and got the same answer in MPB and MPBI:

mpb:
zevenfreqs:, 1, 0, 0, 0, 0, 0, 0.0783938, 0.0784005, 0.172071

mpbi:
zevenfreqs:, 1, 0, 0, 0, 0, 0, 0.0781455, 0.0781697, 0.172016

The slight differences (0.3%, 0.3%, and 0.03%) are due to discretization 
effects and go away quickly as I increase the resolution.  At a resolution 
of 20:

mpb:
zevenfreqs:, 1, 0, 0, 0, 0, 0, 0.0783103, 0.0783116, 0.171905

mpbi:
zevenfreqs:, 1, 0, 0, 0, 0, 0, 0.0782478, 0.0782511, 0.171892

Differences are now 0.08%, 0.08%, and 0.01%, respectively, a reduction by 
about a factor of 4 as one would hope.

Steven