Dear Stano,
Thank you for helping to solve this problem.
As a follow-up, step 2) below should not be required. The qm/mm interface in Amber checks for the Orca installation path and uses the full path to call orca, so it should work in parallel.
All the best,
Andy
—
Dr. Andreas W. Goetz
Assistant Project Scientist
San Diego Supercomputer Center
Tel: +1-858-822-4771
Email: agoetz.sdsc.edu
Web: www.awgoetz.de
> On May 21, 2016, at 12:37 AM, Stanislav Simko <s.simko.uu.nl> wrote:
>
> Dear Collin,
> I've just been dealing with similar problems and for me there are
> several reasons for this:
> 1) there's already another program installed that is called orca
> (screen reader). In this case you just need to do "prepend-path" to
> your orca location, i.e. either (for bash):
> prepend-path PATH "/path/to/orca"
> or smth. like:
> export PATH=path/to/orca:$PATH
>
> 2) Assume that it's installed in the "/apps/orca_directory" directory.
> When running ORCA in parallel you MUST specify whole path to orca
> binary file, i.e.:
> /apps/orca_directory/orca input > output
>
> however, I think this is not the case for the Amber that runs it just
> like "orca" command. Therefore, I suggest you to :
>
> a) create alias for orca command in your .bashrc file, or somewhere
> else:
> alias orca='/full/path/to/orca_dir/orca'
>
> b) add following lines to your submit script so that the alias gets
> interpreted (bash terminal):
> . ~/bashrc
> shopt -s expand_aliases
>
> 3) It may also be that you Orca calculation does not reach the point
> where it creates .engrad file. To test for that I suggest you to just
> take the .inp + coordinates + charges prepared by the Amber and run it
> directly by Orca to find out. It is very probable that you need to add
> section:
> %BASIS GTOAUX=AutoBuild END
>
> to you orc_job.tpl file
>
> Regards,
> stano.
>
>
>
>
>
> On Sat, 2016-05-21 at 05:37 +0000, Nisler, Collin R. wrote:
>> Hello, I am attempting to run a QM/MM simulation with an external
>> interface to Orca. I am able to run it on my machine (albeit very,
>> very slowly). I get the Orca output files and the SCF converges well.
>> However, when I try to run this in parallel (on a supercomputer) I am
>> getting the following error:
>>
>>
>> SANDER BOMB in subroutine read_results (qm2_extern_orc_module)
>> Error reading Orca output from file orc_job.engrad
>> (Current total energy or gradient not found or unsupported units.)
>>
>> I get a similar response when I try to use qchem with amber. I think
>> Amber and Orca are just not able to communicate like they should, but
>> I am completely unsure as to why. This happens both when I submit the
>> job in a batch script, or on the command line. Any suggestions would
>> be greatly appreciated, thanks for your time.
>>
>> Collin Nisler
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon May 23 2016 - 20:00:02 PDT
