Surprising results in the first output of a simulation
Corentin CADIOU
Dear RAMSES User & Developers,
I am getting surprising results when studying angular momentum with RAMSES at very high redshift.
I am studying the evolution of the specific angular momentum of
700 DM halos from a single simulation and I am comparing it to the
prediction from tidal torque theory (TTT). However, there always
seems to be something special happening in the first output
(output_00001) of the simulation, which systematically shows an
excess of angular momentum. This is shown on the figure below
which presents the mean (orange, ±1sigma in shaded orange) and
median (blue, 68% percentiles in shaded blue) of the ratio between
the specific angular momentum of the proto-halos to the specific
angular momentum predicted by TTT, sampled over the entire
population of 700 halos. At these redshifts, the ratio should be
very close to one (since we are far in the linear regime), but as
can be seen on the plot, all halos show a similar excess of
angular momentum in the first output.
The effect seems to be non-physical, as it is only seen in the first
output, regardless of when output_00002 is done (I tried dumping
at z=98,89 and z=15 ; output_00001 is dumped at z=99, and I am
always seeing the effect), so my guess would be that the initial
output is somehow inconsistent with subsequent outputs. However if
it was the case that would probably have been spotted by other
users?
The initial conditions are generated by genetIC (which itself
creates a bunch of MPGrafic files).
Do you have any idea what may be happening?
Regards,
Corentin
-- Dr. Corentin Cadiou Post Doctoral Research Assistant Cosmoparticle Initiative Hub, desk 27 University College London (UCL) Gower St, Bloomsbury, London WC1E 6BT mobile: +33.6.43.18.66.83
Romain Teyssier
Le 23 juin 2020 à 16:50, Corentin CADIOU <corenti...@iap.fr> a écrit :
Dear RAMSES User & Developers,
I am getting surprising results when studying angular momentum with RAMSES at very high redshift.
I am studying the evolution of the specific angular momentum of 700 DM halos from a single simulation and I am comparing it to the prediction from tidal torque theory (TTT). However, there always seems to be something special happening in the first output (output_00001) of the simulation, which systematically shows an excess of angular momentum. This is shown on the figure below which presents the mean (orange, ±1sigma in shaded orange) and median (blue, 68% percentiles in shaded blue) of the ratio between the specific angular momentum of the proto-halos to the specific angular momentum predicted by TTT, sampled over the entire population of 700 halos. At these redshifts, the ratio should be very close to one (since we are far in the linear regime), but as can be seen on the plot, all halos show a similar excess of angular momentum in the first output.
The effect seems to be non-physical, as it is only seen in the first output, regardless of when output_00002 is done (I tried dumping at z=98,89 and z=15 ; output_00001 is dumped at z=99, and I am always seeing the effect), so my guess would be that the initial output is somehow inconsistent with subsequent outputs. However if it was the case that would probably have been spotted by other users?
The initial conditions are generated by genetIC (which itself creates a bunch of MPGrafic files).
Do you have any idea what may be happening?
Regards,
Corentin
<kpejggknclfpdmpp.png>
-- Dr. Corentin Cadiou Post Doctoral Research Assistant Cosmoparticle Initiative Hub, desk 27 University College London (UCL) Gower St, Bloomsbury, London WC1E 6BT mobile: +33.6.43.18.66.83
--
You received this message because you are subscribed to the Google Groups "ramses_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ramses_users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ramses_users/12f51b77-1c34-4b31-8b68-ce60ab15e25f%40iap.fr.
Corentin CADIOU
Hi Romain,
Thanks for your answer. I understand your point that output_00001
reflects whatever the ICs are, but why should there be something
special happening between output_00001 and output_00002 that
doesn't seem to happen between subsequent outputs?
Said differently, if I generate some initial conditions at z=200 and have RAMSES evolve the positions down to z=99 and dump an output at this time, should I expect this output to be contain the same data as if my initial conditions were generated at z=99 (apart from the discrepancies between RAMSES' Poisson solver and whatever time integrator my IC generator uses)?
Corentin
Romain Teyssier
To view this discussion on the web visit https://groups.google.com/d/msgid/ramses_users/8db778e7-c912-a453-bb24-42d443984ab5%40iap.fr.
Corentin CADIOU
Hi Romain and thanks for the answer,
I think that I did not formulate my question clearly. What I was asking was more about the way initial conditions are handled numerically by RAMSES, not about possible physical departure from linearity, so let me reformulate the question, hopefully to make things more clear.
I am wondering whether it is possible for some numerical transients / peculiar features to be present in the first output and only in this one. In my first email, I showed that in my setting, there is a strange evolution of the angular momentum between the first and second output (regardless of the times at which the second output happen). Even if the initial conditions were ill-formed (for example if the initial angular momentum was too high), I would still expect RAMSES numerical integration to follow the linear predictions, especially at very high-z. What I observe instead is that the peculiar behavior always happen between the first and second output, regardless of the time difference between these two.
Thanks for the help,
Corentin