Keeping track of random number

Dear Lammps users
I have a simulation where I use the Langevin thermostat, but for calculations of a certain quantity I need to keep track of the random numbers. Is there an option to output the random numbers to a file somehow or get them by remembering the seed and the reproducing them ?
Although I think implementing this is a compute will be best I just want other options in case I don’t succeed in implementing it
Thank you very much !
Taj

Dear Lammps users
I have a simulation where I use the Langevin thermostat, but for
calculations of a certain quantity I need to keep track of the random

ahh, the infamous undisclosed "certain quantity".

numbers. Is there an option to output the random numbers to a file somehow
or get them by remembering the seed and the reproducing them ?

not without modifying the sources and compiling a custom version and
not without a serious performance penalty.
fix langevin requires a very large number of (pseudo) random numbers.

please also read up in the manual on how to reproduce the sequence of
pRNGs in LAMMPS (or not), particularly for parallel runs.

Hello Axel
Thank you for the quick replay. I am sorry I forget to actually attach the quantity.
I am simulating a system of rigid particles with bonds that can break, under a random potential, This system is embedded in a Langevin bath I wish to calculate the heat flux or rather the fluctuations in the flux form the heat bath into the system as a sanity to monitor jumps
when ever a bond breaks and to get some sort of energetic quantity. The way I thought to do it, is to just calculate the integral over the heat exchange : β ΔQ and for that I need the random numbers.
On another issue, for some reason I cannot access the lammps website from home, I am guessing that my Internet provider is somehow blocked, who should I contact on the list (Some one at Sandia ? )
Again thank you very much for your input.
Taj

Hello Axel
Thank you for the quick replay. I am sorry I forget to actually attach the
quantity.
I am simulating a system of rigid particles with bonds that can break, under
a random potential, This system is embedded in a Langevin bath I wish to
calculate the heat flux or rather the fluctuations in the flux form the heat
bath into the system as a sanity to monitor jumps
when ever a bond breaks and to get some sort of energetic quantity. The way
I thought to do it, is to just calculate the integral over the heat
exchange : β ΔQ and for that I need the random numbers.

huh? i don't follow that argument. what can those random numbers tell
you? why not take that information of the energy transfer directly
from fix langevin?

On another issue, for some reason I cannot access the lammps website from
home, I am guessing that my Internet provider is somehow blocked, who should
I contact on the list (Some one at Sandia ? )

i don't know. i've seen that in the past as well. there is usually a
timeout and then it works again. trying a lot only keeps the block
going. almost all the frequently used information on the webpage is
also part of the source distribution and can be viewed with a
webbrowser locally. not to mention that availability of a pdf file of
the manual.

axel.