Re: [Wien] Problem with x w2w -so -up (wien2wannier)

[Elias Assmann]

On 07/21/2016 09:51 AM, Wen Fong Goh wrote:
> Hi Elias, I have tried out the current development version and it
> reports no error, except it asked for case.inwfup and dn, which I assume
> can be copied from case.inwf. So do you think the current wien2k package
> should be updated with this wien2wannier version?

One more thing: I would advise that you check the results of this
calculation carefully, just to keep an eye on the possibility that the
problem was “fixed” only such the calculation does not crash, but that
some subtler problem remains.

Since it is difficult to check the .amn and .mmn files by eye especially
in the presence of SOC, realistically that means checking the Wannier90
results (even more carefully than you do it anyway …).  If you notice
something strange, please keep us posted.


Elias




signature.asc
Description: OpenPGP digital signature
___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:  
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html

Re: [Wien] Problem with x w2w -so -up (wien2wannier)

[Elias Assmann]

Dear Wen Fong Goh,

Sorry for the delayed reply, and thank you for the report both here
and on GitHub.  Since this problem has become a recurring issue, I will
continue the discussion here, where more users are likely to see it.

On 06/01/2016 01:43 AM, Wen Fong Goh wrote:
> Hi, I know this issue has been floating around in the mailing list, but
> I still couldn't find the solution. When I execute x w2w -up so, I got
> the following error:
> 
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image  PCRoutineLine   
> Source
> w2wc   004EED15  Unknown   Unknown  Unknown
> w2wc   004ECAD7  Unknown   Unknown  Unknown
> w2wc   004946C4  Unknown   Unknown  Unknown
> w2wc   004944D6  Unknown   Unknown  Unknown
> w2wc   0043FED6  Unknown   Unknown  Unknown
> w2wc   00443AF0  Unknown   Unknown  Unknown
> libpthread.so.02EE31850  Unknown   Unknown  Unknown
> w2wc   0046709D  Unknown   Unknown  Unknown
> w2wc   0042B8B3  read_vec_  52 
> read_vec.f

On 06/01/2016 11:57 PM, Wen Fong Goh wrote:
> I don't think that's the case, all of the other work fine with spin
> orbit coupling. What bothers me is that, for the same material, it works
> well with cubic structure, but when I change to tetragonal, it fails.

I am not sure if I reproduced your error.  I do get an error in the
situation you describe, but also in the cubic case.  The messages are
different from what you see as well, but that might be compiler-related.

In the tetragonal case, I see

   *** Error in `w2wc': free(): invalid next size (normal):
0x0235ed90 ***
   forrtl: severe (67): input statement requires too much data, unit 9,
file W-14.2.vectorsoup
   Image  PCRoutineLine
Source

   Stack trace terminated abnormally.

In the cubic case instead it is

*** Error in `w2wc': free(): invalid next size (normal):
0x0257c490 ***
*** Error in `w2wc': free(): invalid next size (normal):
0x025805f0 ***
Abort

Anyway, those errors happen with the version of wien2wannier included
in Wien2k 14.2, which I assume you are using.  With the current
development version, I get no error.

Could you try the development version and report back if the error is
fixed?  If so, could you also check if the results make sense?  I only
tested if w2wc runs without error.

To do this, go to https://github.com/wien2wannier/wien2wannier and
clone or download the repository (there is a green button for that).
When you have that, go to the wien2wannier directory

$ cd wien2wannier-master

and make and unpack a “dist” tarball

$ make wien-dist
$ tar -xvf wien2wannier-*-wiendist.tar.gz

Now I suggest you copy your original Wien2k directory (below, `wien2k')
and replace it with a copy.  This should ensure that the new
executables will be in your path.

$ cp -r wien2k wien2k-new-w2w
$ mv wien2k wien2k-old
$ ln -s wien2k-new-w2w wien2k
$ cd wien2k-new-w2w

There, overwrite the wien2wannier files and recompile

$ tar -xvf
../wien2wannier-master/wien2wannier-*-wiendist/wien2wannier-*-expand-in-wienroot.tar
$ ../wien2wannier-master/compile_wien2wannier.sh

To make sure that you are running the new versions, you can do

$ ls -ld $(dirname $(which w2w))


Elias


PS: Strangely, I have been unable to find your original post in the
archive http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/,
though the answers to it are there.



signature.asc
Description: OpenPGP digital signature
___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:  
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html

Re: [Wien] Problem with x w2w -so -up (wien2wannier)

[Wen Fong Goh]

Hi Laurence, I don't think that's the case, all of the other work fine with 
spin orbit coupling. What bothers me is that, for the same material, it works 
well with cubic structure, but when I change to tetragonal, it fails. Here's 
the struct file in case anyone wants to reproduce the bug. 
https://github.com/wien2wannier/wien2wannier/files/292884/case.struct.txt
Date: Tue, 31 May 2016 19:17:26 -0500
From: laurence.ma...@gmail.com
To: wien@zeus.theochem.tuwien.ac.at
Subject: Re: [Wien] Problem with x w2w -so -up (wien2wannier)

Maybe...the default configuration dies not work if you have a relativistic lo, 
you have to change and recompile.
On May 31, 2016 7:43 PM, "Wen Fong Goh" <christopher_go...@hotmail.com> wrote:






Hi, I know this issue has been floating around in the mailing list, but I still 
couldn't find the solution. When I execute x w2w -up so, I got the following 
error:



forrtl: severe (174): SIGSEGV, segmentation fault occurred

Image  PCRoutineLineSource  
  


w2wc   004EED15  Unknown   Unknown  Unknown

w2wc   004ECAD7  Unknown   Unknown  Unknown

w2wc   004946C4  Unknown   Unknown  Unknown

w2wc   004944D6  Unknown   Unknown  Unknown

w2wc   0043FED6  Unknown   Unknown  Unknown

w2wc   00443AF0  Unknown   Unknown  Unknown

libpthread.so.02EE31850  Unknown   Unknown  Unknown

w2wc   0046709D  Unknown   Unknown  Unknown

w2wc   0042B8B3  read_vec_  52  read_vec.f



The problem is that it works for other cases, but for this particular one it 
fails. Please let me know if you need more information.






___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:  
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html  
 ___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:  
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html

Re: [Wien] Problem with x w2w -so -up (wien2wannier)

[Laurence Marks]

Maybe...the default configuration dies not work if you have a relativistic
lo, you have to change and recompile.
On May 31, 2016 7:43 PM, "Wen Fong Goh" 
wrote:

> Hi, I know this issue has been floating around in the mailing list, but I
> still couldn't find the solution. When I execute x w2w -up so, I got the
> following error:
>
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image  PCRoutineLine
> Source
> w2wc   004EED15  Unknown   Unknown  Unknown
> w2wc   004ECAD7  Unknown   Unknown  Unknown
> w2wc   004946C4  Unknown   Unknown  Unknown
> w2wc   004944D6  Unknown   Unknown  Unknown
> w2wc   0043FED6  Unknown   Unknown  Unknown
> w2wc   00443AF0  Unknown   Unknown  Unknown
> libpthread.so.02EE31850  Unknown   Unknown  Unknown
> w2wc   0046709D  Unknown   Unknown  Unknown
> w2wc   0042B8B3  read_vec_  52
> read_vec.f
>
> The problem is that it works for other cases, but for this particular one
> it fails. Please let me know if you need more information.
>
___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:  
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html