[daip] CALIB

[Crystal Brogan]

WEIGHTIT=1 has made a big improvement.

thanks,
crystal

On Tue, 15 Jul 2008, Eric Greisen wrote:

> Try WEIGHTIT=1
>
> I have been beating myself up trying to figure what has changed if anything 
> in CALIB.  Testing OLD, NEW, and TST do the same with GRID and NEW and TST 
> also do the same with DFT, but version OLD with DFT was okay.  That triggered 
> a memory - DFT weighting was incorrect in modeling and I fixed it.  But CALIB 
> A&P is unstable if weights are correct (outer spacings weighted way low since 
> model ampl way low).  WEIGHTIT was added to stabilize such cases.
>
> The file you sent me does not have an NX table - I need still to find out 
> what troubles that causes.
>
> Cheers,
>
> Eric
>
>