I want to modify the mass of some particles in the TNG snapshot, using the modified snapshot as the initial condition, and re-run a simulation with Gadget4/Arepo code. Is this feasible?
In other words, is the information in the snapshot complete enough that I can start the simulation again?
Best,
Tianning
Dylan Nelson
29 Nov '23
The full snapshots have all the information required to approximately restart a simulation. However, the results will not be identical to having restarted from a set of restartfiles, as the information content is not perfect.
More importantly, since you would be restarting and running with a different physical model, and not the TNG model (?), the results would rapidly diverge from what you see in the TNG snapshots.
Tianning LYU
30 Nov '23
Dear Dylan,
Thank you for your reply. But if I were to use data from pure-DM simulations (such as TNG50-1-Dark), would the physical model still have such a significant impact?
My idea was to investigate whether a change in the mass of a particle would cause a large change in its spatial position and kinematics in the halo, i.e., assuming that my changes would only affect the kinematics of that particle, rather than the overall simulation results. If the simulation results are significantly different from TNG, my idea will not be realized, so that I need to think about other ideas to achieve my goal.
Best,
Tianning
Dylan Nelson
30 Nov '23
Hi Tianning,
A DM-only simulation makes a lot more sense here, I see.
I think you should just give it a try: first, just restarting from a snapshot, and not making any changes. You can see how close to the "original answer" (subsequent TNG snapshots) your simulation result is.
Tianning LYU
30 Nov '23
OK I will have a try. Thank you for the timely and careful reply!
Dear all,
I want to modify the mass of some particles in the TNG snapshot, using the modified snapshot as the initial condition, and re-run a simulation with Gadget4/Arepo code. Is this feasible?
In other words, is the information in the snapshot complete enough that I can start the simulation again?
Best,
Tianning
The full snapshots have all the information required to approximately restart a simulation. However, the results will not be identical to having restarted from a set of restartfiles, as the information content is not perfect.
More importantly, since you would be restarting and running with a different physical model, and not the TNG model (?), the results would rapidly diverge from what you see in the TNG snapshots.
Dear Dylan,
Thank you for your reply. But if I were to use data from pure-DM simulations (such as TNG50-1-Dark), would the physical model still have such a significant impact?
My idea was to investigate whether a change in the mass of a particle would cause a large change in its spatial position and kinematics in the halo, i.e., assuming that my changes would only affect the kinematics of that particle, rather than the overall simulation results. If the simulation results are significantly different from TNG, my idea will not be realized, so that I need to think about other ideas to achieve my goal.
Best,
Tianning
Hi Tianning,
A DM-only simulation makes a lot more sense here, I see.
I think you should just give it a try: first, just restarting from a snapshot, and not making any changes. You can see how close to the "original answer" (subsequent TNG snapshots) your simulation result is.
OK I will have a try. Thank you for the timely and careful reply!