[Wien] e-test: subscript out of range

[Stefaan Cottenier]

> I just installed Mandrake Linux and wien2k without any problems,
> but after the third iteration of the calculations wien stops with the
> error message
>
> e-test: subscript out of range
>
> Has anyone an idea?

I'm very happy to see this message [sorry :-) ], as finally somebody might
have encountered a problem similar to the one that is bothering me (and
apparantly only me) for years. On Athlon + Suse (and only there, no problems
on P-IV + Suse), the first two iterations of any scf-cycle stop with the
following message (see bottom line, there you see your e-test message) :

(...)
FORTRAN STOP  CORE  END
FORTRAN STOP  MIXER END
(standard_in) 1: parse error
(standard_in) 1: illegal character: :
(standard_in) 1: parse error
(standard_in) 1: illegal character: :
(standard_in) 1: illegal character: :
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: illegal character: :
(standard_in) 1: parse error
(standard_in) 1: illegal character: :
(standard_in) 1: illegal character: :
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
(standard_in) 1: parse error
etest: Subscript out of range.

This problem disappears as soon as the 3th iteration is started. Therefore,
I have by default to run an scf-cycle by

run_lapw -i 1
run_lapw -i 1
run_lapw -cc 0.00001 etc.,

i.e. I have to add two single iterations before the actual run_lapw. Not
harmful, but boring.

If your problem is related to mine, it might disappear if you just continue
(you have the problem starting from the 3th iteration, for me it disappears
there, our situations might be different hence).

Is there anybody who has similar problems, or who knows how to overcome this
in an elegant way ?

Stefaan