bader code gives segmentation fault with option 1,2 and 3

Bader charge density analysis

Moderator: moderators

Post Reply
amit

bader code gives segmentation fault with option 1,2 and 3

Post by amit »

I tried downloading the binary for linux and also compiling the source code given on the webpage, but each time I tried giving options 1,2 or 3 it gives segmentation fault even with the examples ofr NaCl. Please advise
thanks
andri
Site Admin
Posts: 117
Joined: Tue Apr 26, 2005 6:20 am

Post by andri »

Can you give a little bit more information, e.g. type and version of the linux distribution and type and version of the compiler you're using. Also maybe send the first lines of your CHGCAR or cube file, at least enough to include of few lines of the charge grid. And anything else you can think of that could be helpful.

Thanks!
amit

Re:

Post by amit »

I tried running in 2 systems :
(i) custom compiled SMP Linux kernel patched by openMOSIX 2.4.20-2, on a RedHat 9.0 (Shrike) base (intel fortron compiler version 8.0)
(ii) SMP Linux kernel 2.4.21, on a Scientific Linux 3.0.4-x86_64 (intel fortron compiler version 8.1)
The CHGCAR given in the webpage for NaCl
( http://theory.cm.utexas.edu/henkelman/r ... 160.tar.gz)
was unzipped and used.
I also tried using the binary for linux from your webpage (http://theory.cm.utexas.edu/henkelman/r ... lnx.tar.gz).
The code works good for Option 4 but when Options 1,2 or 3 are given it just gives ACF.dat, AtomVolumes.dat, BCF.dat, CHGCAR, core.3143 dipole.dat and along with this there is an error :Segmentation Fault (Core dumped)
Hope this explains everything. Please let me know if there is anything more you need to know.
Thanks
andri
Site Admin
Posts: 117
Joined: Tue Apr 26, 2005 6:20 am

Post by andri »

This seems to me to have more to do with ifort than the code itself. All other version (i.e. for Mac and Win) run as expected. I tried to compile it on linux with lf95 and pgf90 and they both run fine. So I have added a link on the web site for a executable compiled with pgf90. The source file directory includes the Makefile as well. Hopefully it will work for you now!
amit

Re:

Post by amit »

Thanks a lot for your time. I will surely try with the binary, but you may be correct or it may be due to some other problem. Today I was analysing the CHGCAR files of some relaxed structures with 20 atoms per unit cell. In some of structures the code ran well for OPTION 4 but in 2 cases it reported segmentation fault. I dont think its due to RAM as we have 8GB RAM per node and also the CHGCAR file looks good and I converted it to xsf format and viewed it through XCRYSDEN. Will report more about results with the new binary.
amit

Re:

Post by amit »

I have posted my CHGCAR and CHGCARxsf files in the following link
http://mse-uc01.eng.ohio-state.edu/samanta/delete_Ni/
Amit

Re:

Post by Amit »

Yes the new binary works well. Thanks a lot !!!
andri
Site Admin
Posts: 117
Joined: Tue Apr 26, 2005 6:20 am

Post by andri »

Excellent news. File free to contribute any useful scripts to the site. It as long been a problem how to visualize the CHGCAR files. How does this XCRYSDEN perform?

Thanks!
amit

xcrysden

Post by amit »

I think xcrysden is pretty good, though slow. We have a small code that converts CHGCAR or any other files in similar format to XSF format (http://164.107.79.177/Archive/Graphics/A/utils.html , look for chg2xsf ). Infact there is another code which is hooked to XCRYSDEN webpage written by Jens Kunstmann (http://www.xcrysden.org/Description.html or http://www.fkf.mpg.de/andersen/users/ku ... tware.html ). There are few more codes for VASP postprocessing and a very fast visualizer ATOMEYE that we use for general visualization purposes.
I was wondering if there is any way by which we could have some kind of visualization of the strucutres with atoms surrounded by small regions denoting BADER VOLUME.
andri
Site Admin
Posts: 117
Joined: Tue Apr 26, 2005 6:20 am

Post by andri »

Thanks for the links. This all looks very interesting. We are Mac OS based here and I tried the executable for AtomEye on my machine but it failed. Is there any way to get an updated version or even possibly the source code?

Have you ever tried to use VASPVIEW, http://vaspview.sourceforge.net/ ? It can read VASP CHGCAR files and the volumes files written out by the Bader program have the same format.

Andri
amit

atomeye

Post by amit »

Atomeye for Mac is only available for OS X or latter versions. We just now checked the binary on OS X with kernel version 8.2.0 and it works pretty well. Either way round you can find the source code in this link (http://164.107.79.177/Archive/Graphics/ ... 915.tar.gz )
Hope that helps.
andri
Site Admin
Posts: 117
Joined: Tue Apr 26, 2005 6:20 am

Post by andri »

I'm running 10.3.9 with kernel 7.9.0 and the executable didn't work for me as previously mentioned. I did try it on a machine running 10.4.x and it worked fine there. Also I tried to compile my own executable and managed after some struggle (basically had to install some libraries ) to compile all the subparts of the program but it wouldn't link together, it complained about some library issues, something about the libreadline.a. Is there a special version of gcc or something that you guys recommend?
amit

re

Post by amit »

I am unable to suggest anything on that. All our systems are linux based and hence I dont know very much about Mac. I am sorry for that. But you said it works in one of the machines hopefully you can use that .
Post Reply