[gmx-users] what's the difference between gen_seed and ld_seed?

[Albert]

hello :

I am a little bit confused about the difference between gen_seed and
ld_seed. I checked the manual, it is said:

gen_seed
used to initialize random generator for random velocities, when gen_seed
is set to -1, the seed is calculated from the process ID number.
This is often used coupled with gen_vel which is Generate velocities in
grompp according to a Maxwell distribution at temperaturegen_temp [K],
with random seed gen_seed. This is only meaningful with integratormd.

As indicated in Jonhn E.Kerrigan's tutorial, gen_seed=-1 is always
turned on in NVT, NPT and MD production step. However, in Justin's
Lysozyme in Water tutorial, this option is only turned on in NVT and the
following NPT and MD production were turned off.

I am just wondering which option would be better for our simulations?

How about ld_seed? here is the statement from manual:

ld_seed: (1993) [integer]
used to initialize random generator for thermal noise for stochastic and
Brownian dynamics. When ld_seed is set to -1, the seed is calculated
from the process ID. When running BD or SD on multiple processors, each
processor uses a seed equal to ld_seed plus the processor number.

when should we turn this option on?

thank you very much
best
Albert

[Justin Lemkul]

I am a little bit confused about the difference between gen_seed and ld_seed.
gen_seed
used to initialize random generator for random velocities, when gen_seed is set
to -1, the seed is calculated from the process ID number.
This is often used coupled with gen_vel which is Generate velocities in grompp
according to a Maxwell distribution at temperaturegen_temp [K], with random seed
gen_seed. This is only meaningful with integratormd.
As indicated in Jonhn E.Kerrigan's tutorial, gen_seed=-1 is always turned on in
NVT, NPT and MD production step. However, in Justin's Lysozyme in Water
tutorial, this option is only turned on in NVT and the following NPT and MD
production were turned off.
I am just wondering which option would be better for our simulations?

Regenerating velocities at each step negates the purpose of previous
equilibration. You should generate velocities only once. Note that gen_seed is
ignored if gen_vel = no.

ld_seed: (1993) [integer]
used to initialize random generator for thermal noise for stochastic and
Brownian dynamics. When ld_seed is set to -1, the seed is calculated from the
process ID. When running BD or SD on multiple processors, each processor uses a
seed equal to ld_seed plus the processor number.
when should we turn this option on?

Which integrator are you using? For the md integrator, ld_seed is irrelevant.
For sd and bd, it is pertinent.

-Justin
--
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | [phone]
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================

[Albert]

hello guys

I am producing MD production. That's really helpful.

thank you very much
best
Albert

Post by Justin Lemkul

I am a little bit confused about the difference between gen_seed and ld_seed.
gen_seed
used to initialize random generator for random velocities, when gen_seed is set
to -1, the seed is calculated from the process ID number.
This is often used coupled with gen_vel which is Generate velocities in grompp
according to a Maxwell distribution at temperaturegen_temp [K], with random seed
gen_seed. This is only meaningful with integratormd.
As indicated in Jonhn E.Kerrigan's tutorial, gen_seed=-1 is always turned on in
NVT, NPT and MD production step. However, in Justin's Lysozyme in Water
tutorial, this option is only turned on in NVT and the following NPT and MD
production were turned off.
I am just wondering which option would be better for our simulations?

Regenerating velocities at each step negates the purpose of previous
equilibration. You should generate velocities only once. Note that
gen_seed is ignored if gen_vel = no.

ld_seed: (1993) [integer]
used to initialize random generator for thermal noise for stochastic and
Brownian dynamics. When ld_seed is set to -1, the seed is calculated from the
process ID. When running BD or SD on multiple processors, each processor uses a
seed equal to ld_seed plus the processor number.
when should we turn this option on?

Which integrator are you using? For the md integrator, ld_seed is
irrelevant. For sd and bd, it is pertinent.
-Justin

[Peter C. Lai]

Post by Albert
I am a little bit confused about the difference between gen_seed and
gen_seed
used to initialize random generator for random velocities, when gen_seed
is set to -1, the seed is calculated from the process ID number.
This is often used coupled with gen_vel which is Generate velocities in
grompp according to a Maxwell distribution at temperaturegen_temp [K],
with random seed gen_seed. This is only meaningful with integratormd.
As indicated in Jonhn E.Kerrigan's tutorial, gen_seed=-1 is always
turned on in NVT, NPT and MD production step. However, in Justin's
Lysozyme in Water tutorial, this option is only turned on in NVT and the
following NPT and MD production were turned off.

Yes because you undo your equilibration when you reassign initial
velocities using gen_seed in NPT and MD production (NPT) (md/md-vv
integrator).

Post by Albert
I am just wondering which option would be better for our simulations?
ld_seed: (1993) [integer]
used to initialize random generator for thermal noise for stochastic and
Brownian dynamics. When ld_seed is set to -1, the seed is calculated
from the process ID. When running BD or SD on multiple processors, each
processor uses a seed equal to ld_seed plus the processor number.
when should we turn this option on?

As stated, you use this option for use with the bd or sd integrators.

[Thomas Schlesier]

Post by Peter C. Lai

Post by Albert
I am a little bit confused about the difference between gen_seed and
gen_seed
used to initialize random generator for random velocities, when gen_seed
is set to -1, the seed is calculated from the process ID number.
This is often used coupled with gen_vel which is Generate velocities in
grompp according to a Maxwell distribution at temperaturegen_temp [K],
with random seed gen_seed. This is only meaningful with integratormd.
As indicated in Jonhn E.Kerrigan's tutorial, gen_seed=-1 is always
turned on in NVT, NPT and MD production step. However, in Justin's
Lysozyme in Water tutorial, this option is only turned on in NVT and the
following NPT and MD production were turned off.

Yes because you undo your equilibration when you reassign initial
velocities using gen_seed in NPT and MD production (NPT) (md/md-vv
integrator).

Post by Albert
I am just wondering which option would be better for our simulations?
ld_seed: (1993) [integer]
used to initialize random generator for thermal noise for stochastic and
Brownian dynamics. When ld_seed is set to -1, the seed is calculated
from the process ID. When running BD or SD on multiple processors, each
processor uses a seed equal to ld_seed plus the processor number.
when should we turn this option on?

As stated, you use this option for use with the bd or sd integrators.

Also the 'v-rescale' thermostat uses the ld_seed.