Re: [AMBER] parallelization of paramfit in ambertools12

From: Francesco Pietra <chiendarret.gmail.com>
Date: Tue, 17 Apr 2012 07:28:02 +0200

Hi Robin:
That helps much. I am only interested in paramfit, so that the second
option is what I want. However, is it any configuration to parallelize
paramfit on more that two processors?

More importantly, is it any hope to adjust vdw params for atoms that
are not comprised in antechamber? I mean that I plan to choose by
intuition such params and hope that paramfit will correct them
appropriately.

thanks

(rbetz.ucsd is not recognized by my provider)

francesco pietra

On Tue, Apr 17, 2012 at 12:13 AM, Robin Betz <rbetz.ucsd.edu> wrote:
> Hi Francesco,
>
> You need to re-run the configure script in $AMBERHOME with the -openmp
> option:
> $AMBERHOME/configure -openmp gnu (or intel if you are using those
> compilers).
>
> To only recompile paramfit with the openmp, go to
> $AMBERHOME/AmberTools/src/paramfit
> and run make install
> Paramfit's header message will have a line saying  something like:
>
> *************************************************************************************
>                                Execution started at:
> |                             Mon Apr 16 15:11:56 2012
> |
> |                 Running OpenMP version of code using  2 processors
>
> if it has been compiled correctly with OpenMP support.
>
> Hope this helps,
>
> Robin Betz
>
>
>
>
>
> On Mon, Apr 16, 2012 at 12:47 AM, Francesco Pietra <chiendarret.gmail.com>wrote:
>
>> Hello:
>> I have installed without errors (after commenting out line 450 of
>> "configure2" as Prof Case suggested)
>> serial-ambertools12/serial-amber10 in Debian amd64 in a game machine
>> with six CPU's and two GTX-580 cards. I would be much interested in
>> having "paramfit" parallelized. I have an installation of OpenMPI.
>>
>> It is not clear to me when and where "Enable this by adding the
>> -openmp option to configure and rebuilding paramfit" should be done. I
>> can imagine on this particular code within ambertools12,nonetheless a
>> bit of guideline would be helpful.
>>
>> Thanks
>>
>> francesco pietra
>>
>> _______________________________________________
>> 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 Apr 16 2012 - 22:30:03 PDT
Custom Search