[gmx-users] What are the random factors in mdrun?

[Mark Abraham]

Justin A. Lemkul wrote:
> 
> 
> Lee Soin wrote:
>> Hello!
>>  
>> I am confused about random factors in GROMACS, because I find that 
>> different runs of the same .tpr file do not generate identical 
>> results. I know that Langevin dynamics introduces random forces, but 
>> what are the other random factors in GROMACS? Many thanks in advance.
>>
> 
> There is some useful information here:
> 
> http://oldwiki.gromacs.org/index.php/Extending_Simulations
> 
> Check the "Exact versus binary identical continuation" section.

To achieve reproducibility from the same .tpr file requires that you use 
"mdrun -reprod". There have been discussions on this list in the last 
month on this topic.

Mark