Dear All,
Is it known if there are any issues with the VTST tools and vasp 5.2.11 (latest release as of December 2010) ?
All advice gratefully received.
Iain
vasp 5.2.11
Moderator: moderators
Re: vasp 5.2.11
There are some build issues unrelated to the vtstcode. These involve some missing use module statements in the code. We have not seen any changes that interact with the vtstcode.
There could be issues with any xdatcar related scripts. The release notes say that there is a change in the header, but we have not seen any problems yet.
There could be issues with any xdatcar related scripts. The release notes say that there is a change in the header, but we have not seen any problems yet.
Re: vasp 5.2.11
Thanks for the swift reply.
If you know which modules need to be modified to get a successful build, that would be very helpful.
I have a successful build of 5.2.11 on SciNet (IBM AIX power6 cluster with imbs mpxlf90 compiler)
but I have not been able to make a build using VTST modules.
My build dies at
mpxlf90 -qfree=f90 -qmaxmem=-1 -qarch=auto -qtune=auto -qcache=auto -qinitauto -qcheck -qsave=all -O3 -qarch=auto -qstrict -c hessian.f
"hessian.f", line 22.30: 1514-084 (S) Identifier io is being declared with type name in_struct which has not been defined in a derived type definition.
** hessian === End of Compilation 1 ===
1501-511 Compilation failed for file hessian.f.
although in the absence of the VTST modules, hessian compiles without this message.
All help would be very gratefully received. We do appreciate that you do this out of kindness!
best regards... Iain McNab
If you know which modules need to be modified to get a successful build, that would be very helpful.
I have a successful build of 5.2.11 on SciNet (IBM AIX power6 cluster with imbs mpxlf90 compiler)
but I have not been able to make a build using VTST modules.
My build dies at
mpxlf90 -qfree=f90 -qmaxmem=-1 -qarch=auto -qtune=auto -qcache=auto -qinitauto -qcheck -qsave=all -O3 -qarch=auto -qstrict -c hessian.f
"hessian.f", line 22.30: 1514-084 (S) Identifier io is being declared with type name in_struct which has not been defined in a derived type definition.
** hessian === End of Compilation 1 ===
1501-511 Compilation failed for file hessian.f.
although in the absence of the VTST modules, hessian compiles without this message.
All help would be very gratefully received. We do appreciate that you do this out of kindness!
best regards... Iain McNab
Re: vasp 5.2.11
Sure, the changes are to add
USE main_mpi
any place that you see errors about DIR_APP or DIR_LEN, such as dynconstr.F, dimer_heyden.F and gadget.F
and
USE base
to use in_struct in hessian.F
I also had to change the name of a lapack routine that was not included in the vasp lapack library. I can't remember the name, but it was specific to scalapack and I switched it to the equivalent regular lapack function.
USE main_mpi
any place that you see errors about DIR_APP or DIR_LEN, such as dynconstr.F, dimer_heyden.F and gadget.F
and
USE base
to use in_struct in hessian.F
I also had to change the name of a lapack routine that was not included in the vasp lapack library. I can't remember the name, but it was specific to scalapack and I switched it to the equivalent regular lapack function.
Re: vasp 5.2.11
Dear Graeme,
That's wonderful. I was now able to build VASP.5.2.11 with the vtst codes.
Thank you very much for your help.
best regards... Iain
That's wonderful. I was now able to build VASP.5.2.11 with the vtst codes.
Thank you very much for your help.
best regards... Iain
Re: vasp 5.2.11
the scalapack function is dlasrt2 in david_inner.F and can be switch with the lapack version dlasrt