[Wien] lapwdm angle.f

[Peter Blaha]

Thank's for the hint for angle.f in SRC_lapwdm. Yes, I think the same fix should
be applied to this program. (Apparently both routines did not support a 
triclinic lattice).

The "logic" of these "if statements" should be such that

the first case handles a monoclinic lattice with angle gamma .ne. 90 (and the 
others are 90)

the second with beta.ne.90

the third case handles a triclinic lattice.

And yes, you are absolutely right, in the fix is a stupid misstype. It should 
be:

  ELSE IF 
((ABS(ALPHA(2)-PI/2.d0).GT.1.D-4).and.(ABS(ALPHA(1)-PI/2.d0).LT.1.D-4).and.(ABS(ALPHA(3)-PI/2.d0).LT.1.D-4))
 THEN

^

PS: I'm not sure if all this is really necessary or one could stick to the last 
general formulae.

PPS: I'd like to thank you explicitly for your many valuable comments to the 
mailing list in the last months.
  Thank you very much !!!



Am 02.07.2012 00:56, schrieb Gavin Abo:
> Dear Dr. Blaha and developers,
>
> Fixes to angle.f in SRC_qtl was reported:
>
> http://zeus.theochem.tuwien.ac.at/pipermail/wien/2012-June/017085.html
>
> It suggested changes to the if statements.
> *
> IF (ABS(ALPHA(3)-PI/2.d0).GT.1.D-4) THEN*
>
> to
>
> *IF 
> ((ABS(ALPHA(3)-PI/2.d0).GT.1.D-4).and.(ABS(ALPHA(2)-PI/2.d0).LT.1.D-4).and.(ABS(ALPHA(1)-PI/2.d0).LT.1.D-4))
>  THEN*
>
> and
>
> *ELSE IF (ABS(ALPHA(2)-PI/2.d0).GT.1.D-4) THEN*
>
> to
>
> *ELSE IF 
> ((ABS(ALPHA(2)-PI/2.d0).GT.1.D-4).and.(ABS(ALPHA(1)-PI/2.d0).LT.1.D-4).and.(ABS(ALPHA(2)-PI/2.d0).LT.1.D-4))
>  THEN*
>
> The angle.f in SRC_lapwdm is almost the same.  Should these fixes be applied 
> to it too?
>
> Can you confirm that the absolute values such as on 
> "ABS(ALPHA(3)-PI/2.d0).GT.1.D-4" are correct? It may be correct, but it looks 
> suspicious to me.  The left side argument will
> always be positive because of the absolute value.  Therefore, the if 
> statements may never have a different result (true or false) unless the left 
> argument is almost zero.
>
> P.S., Kateryna, your ORBxxx and SPIxxx results might change if a fix is 
> needed.  However, that is only if the code needs to select a different set of 
> crystal geometry equations in
> angle.f for your crystal structure instead of:
>
>  IF (ABS(ALPHA(3)-PI/2.d0).GT.1.D-4) THEN
>  XA=XMS(1)*AA*SIN(ALPHA(3))
>  XB=XMS(1)*AA*COS(ALPHA(3))+BB*XMS(2)
>  XC=CC*XMS(3)
>
> Kind Regards,
>
> Gavin Abo
>
>
> ___
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>

-- 

   P.Blaha
--
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300 FAX: +43-1-58801-165982
Email: blaha at theochem.tuwien.ac.atWWW: http://info.tuwien.ac.at/theochem/
--


-- next part --
A non-text attachment was scrubbed...
Name: angle.f
Type: text/x-fortran
Size: 2475 bytes
Desc: not available
URL: 

[Wien] lapwdm angle.f

[Gavin Abo]

Dear Dr. Blaha and developers,

Fixes to angle.f in SRC_qtl was reported:

http://zeus.theochem.tuwien.ac.at/pipermail/wien/2012-June/017085.html

It suggested changes to the if statements.
*
IF (ABS(ALPHA(3)-PI/2.d0).GT.1.D-4) THEN*

to

*IF 
((ABS(ALPHA(3)-PI/2.d0).GT.1.D-4).and.(ABS(ALPHA(2)-PI/2.d0).LT.1.D-4).and.(ABS(ALPHA(1)-PI/2.d0).LT.1.D-4))
 
THEN*

and

*ELSE IF (ABS(ALPHA(2)-PI/2.d0).GT.1.D-4) THEN*

to

*ELSE IF 
((ABS(ALPHA(2)-PI/2.d0).GT.1.D-4).and.(ABS(ALPHA(1)-PI/2.d0).LT.1.D-4).and.(ABS(ALPHA(2)-PI/2.d0).LT.1.D-4))
 
THEN*

The angle.f in SRC_lapwdm is almost the same.  Should these fixes be 
applied to it too?

Can you confirm that the absolute values such as on 
"ABS(ALPHA(3)-PI/2.d0).GT.1.D-4" are correct? It may be correct, but it 
looks suspicious to me.  The left side argument will always be positive 
because of the absolute value.  Therefore, the if statements may never 
have a different result (true or false) unless the left argument is 
almost zero.

P.S., Kateryna, your ORBxxx and SPIxxx results might change if a fix is 
needed.  However, that is only if the code needs to select a different 
set of crystal geometry equations in angle.f for your crystal structure 
instead of:

 IF (ABS(ALPHA(3)-PI/2.d0).GT.1.D-4) THEN
 XA=XMS(1)*AA*SIN(ALPHA(3))
 XB=XMS(1)*AA*COS(ALPHA(3))+BB*XMS(2)
 XC=CC*XMS(3)

Kind Regards,

Gavin Abo
-- next part --
An HTML attachment was scrubbed...
URL: