On 8/29/2024 3:27 PM, David A Case via AMBER wrote:
> On Wed, Aug 28, 2024, Tamalika Ash via AMBER wrote:
>>
>> *Note: The following floating-point exceptions are signalling:
>> IEEE_INVALID_FLAG IEEE_UNDERFLOW_FLAG IEEE_DENORMALERROR: Calculation
>> halted. Periodic box dimensions have changed too much from their
>> initial
>> values.*
>> *Your system density has likely changed by a large amount, probably from
>> starting the simulation from a structure a long way from equilibrium.*''
>
> In addition to Tim's suggestion of just restarting, you might consider
> this:
> run your initial equilibration (just long enough to get the system volume
> stabilized) using the CPU version. Then you can switch to the GPU.
>
> ...good luck...dac
>
Let me add to Dave and Tim here. Equilibrating solvent simulation
cells, especially at the start of simulations is an important first
step. Here are some general recommendations that you may want to
consider (they are subjective, or course).
You should be able to make a reasonable estimate of the target density
(typically 1.0-1.2 g/mL is reasonable) of your system, and hence what a
reasonable volume should be. In some cases you may also want to enforce
a specific shape for the cell (for example, if you want to prepare a
system so as to be able to be connected to another simulations with
Hamiltonian replica exchange - the unit cell vectors will need to be the
same within an isotropic scaling constant).
We have implemented a simple flag that makes these two things easier to
achieve. You can use ntp = 4 to slowly force the system to have a
specific set of lattice vectors over a given number of steps. These
parameters are controlled by the cntrl name list target_a, target_b,
target_c (the target a,b and c lattice vector lengths) and target_n in
the ewald namelist for number of volume iterations to reach the target
volume respectively.
See "21.6.8. Pressure regulation" of the Amber24 manual for further details.
Once you force the system to roughly to the shape and volume you want,
then you can minimize the system (including the lattice vectors) and
then start dynamics again under constant pressure heating the system
slowly back to the target temperature.
While you are doing all this, you usually want to restrain your solute
so that it does not adopt an artificial conformation - don't let it have
conformational freedom until the solvent settles down and equilibrates.
I hope you find these suggestions useful.
best wishes,
Darrin
--
======================================================================
Darrin M. York : Distinguished Professor
: Henry Rutgers University Chair
Institute for Quantitative :
Biomedicine and Department of : Director, Laboratory for
Chemistry & Chemical Biology : Biomolecular Simulation Research
:
Rutgers, the State University : Darrin.York.rutgers.edu
of New Jersey : fax: +1-732-445-4320
174 Frelinghuysen Road : phone: +1-848-445-5199
Piscataway, NJ 08854 USA : http://theory.rutgers.edu
======================================================================
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 30 2024 - 07:30:02 PDT
