Hi Thorsten,
Yes, X(I) is the optimized parameter, and, if I remember correctly,
D(I) is some (method-dependent) scaling parameter, and G(I) is the
magnitude of the gradient. The stop tolerances and other parameters
can be changed, which is described very briefly in the manual (see the
table of internal variables). But if you want more information, you'll
have to look up the original papers on this optimization routine (Refs
73 & 74 in the spinev JMR paper).
Best,
Mikhail