UPF Pseudopotential

Tim · Post by **Tim** » Wed Jul 06, 2011 4:42 am

Does ABINIT support UPF pseudopotentials? Based on another forum post I found, it appears that it does. And when I run ABINIT (6.8.1) with a UPF pseudopotential it does not complain. However I'm seeing a strange bug. When I run a calculation on Zn with ecut=20, I get the following error in the output:

Code: Select all

- pspatm: opening atomic psp file    Zn_HSCV_PBE-1.0.upf
  Reading pseudopotential file in UPF format...
  ...done
 upf2abinit: XC string from pseudopotential is :
> XXX  CCC  XXX CCC  <

 pspatm: WARNING -
  Pseudopotential file pspxc=       2,
  not equal to input ixc=      11.
  These parameters must agree to get the same xc 
  in ABINIT code as in psp construction.
  Action : check psp design or input file.
  Assume experienced user. Execution will continue.

 pspatm: epsatm=   63.12654446
         --- l  ekb(1:nproj) -->
             1   -0.022489
forrtl: severe (27): too many records in I/O statement, unit -5, file Internal Formatted Write
Image              PC                Routine            Line        Source             
libirc.so          00002B1931D0E396  Unknown               Unknown  Unknown
libirc.so          00002B1931D0D328  Unknown               Unknown  Unknown
libifcoremt.so.5   00002B193169C736  Unknown               Unknown  Unknown
libifcoremt.so.5   00002B19316236A4  Unknown               Unknown  Unknown
libifcoremt.so.5   00002B1931622F06  Unknown               Unknown  Unknown
libifcoremt.so.5   00002B193166C17C  Unknown               Unknown  Unknown
abinit-serial      0000000000F7DFCE  Unknown               Unknown  Unknown
abinit-serial      0000000000F69E78  Unknown               Unknown  Unknown
abinit-serial      0000000000510885  Unknown               Unknown  Unknown
abinit-serial      0000000000468517  Unknown               Unknown  Unknown
abinit-serial      000000000045C8B2  Unknown               Unknown  Unknown
abinit-serial      0000000000453AFA  Unknown               Unknown  Unknown
abinit-serial      0000000000452B92  Unknown               Unknown  Unknown
libc.so.6          0000003904C1C3FB  Unknown               Unknown  Unknown
abinit-serial      0000000000452ADA  Unknown               Unknown  Unknown

When I use ecut=40, the calculation appears to run fine, except I still get this in the output:

Code: Select all

 pspatm: epsatm=   63.12654446
         --- l  ekb(1:nproj) -->
             1   -0.022489
             2

Is the last line supposed to be blank after the 2? I can't seem to find an obvious trend for why this error shows up with some values of ecut and not others. There is some evidence that the error is less likely to appear for high values of ecut.