[Wien] Some questions about the vec2ascii program

[Elias Assmann]

Hi Majid,

I am cross-posting this to the mailing list.

On 10/25/2014 12:34 PM, Majid Yazdani wrote:
> I have two questions about the vec2ascii program. Could you help me to
> solve them?

First off, I do not really consider ‘vec2ascii’ fully supported.  I 
wrote that code for a specific problem and thought it might be helpful 
for other people as well.  I have not tested it since then.

> I use the wien2k14.1 code to calculate the electronic and optical
> properties of a compound by 12 core Intel processor.I need to change the
> case.vectores for some calculations. So I use the vec2ascii which is
> implemented in the wien2k14.1 to do this.

If you want to “change the .vector files” (to continue with lapw2 using 
the new ones?) are you sure vec2ascii is the best way to proceed?  If 
you need to write them in binary, maybe you can skip the plain-text step.

> After running this program a case.vector_ascii is produced.

That's already good news, I suppose!  ;-)

> The size of the case. vector_ascii is:
>
> [papi at cm6 case]$ du -sh case.vector_ascii
>
> 23G     case.vector_ascii
>
> While the size of the case.vectore is:
>
> [papi at cm6 case]$ du -sh case.vector*
>
> 619M    case.vector_1
…
> 620M    case.vector_12

> Why is the size of the case. vector_ascii very larger than that of  the
> case.vectors?

Because the regular vector files are in binary rather than plain text. 
Thus they need much less space and less time to read and write.  I 
suppose that is the reason why they are stored in binary in the first 
place.  (You can try zipping the plain-text file, that should get it 
back to roughly the binary-file size.)

> Is there any program to reproduce the case.vectors from the modified
> case.vector_ascii file so that they are suitable for the Wien2k code?

Not that I know of.  Of course, you can try “reversing” the input code …


-- 
Elias Assmann
Institute of Solid State Physics
Vienna University of Technology
<http://www.ifp.tuwien.ac.at/cms/>