I hope you are doing well. I'm looking for something called the 'parameter_file' for IllustrisTNG, which is used to plan and create compound light rays joining up multiple snapshots.
The yt function I'm trying to use is here. The file itself is (I think) described better in trident documentation, in the fourth paragraph after the note here.
I've seen that you provide initial conditions files in .HDF5 format. However my attempt to use this hasn't worked so far.
Please could you let me know whether this is indeed the file I should be using, and if not, where I can find the correct one?
I am not sure what the contents of this parameter file should be, I think you will have to ask the yt/trident developers (if not clear from the documentation or code), and then construct this file.
Thanks very much Dylan! I shall do so.
The yt developers have pointed me in the right direction. They suggest that I should be able to use this function, and speculate that the way to do so is to create a parameter file for IllustrisTNG based on a Gadget template, as this is similar to Arepo. They've provided me with a Gadget template which works, which is a text file which I've uploaded to github here.
It contains some very specific information, by the looks of it. Would you be ok with me having the info, and if so, please could you help me put together the file?
I'm afraid I cannot construct such a file for you, this could be some work and I wouldn't want to make mistakes. You can of course set the cosmology correctly, and the unit system for: length, mass, and velocity are correct as is. Other than that, many of these options do not have direct analogs in AREPO, and/or may be misleading if adopted directly, e.g. the gravitational softening of gas cannot be described in the way of this parameter file.
Thus I would suggest to set these simple things and get it a try. You will have to assess whether or not the outputs make sense, i.e. know which parameters are being used by the code to be confident they are being used correctly.
Thanks for the advice. No worries, I'll keep consulting with the yt developers and try putting a file together myself and test it. I shall let you know how it goes, if you wish.
I'll chime in here--I think that the only parameters that Charlie needs are the ones which define units, the cosmological parameters, and the redshift/scale factor cadence at which snapshots are saved. He shouldn't need anything else like the smoothing length or anything like that.
Hi Dylan and John,
Thanks to John and Britton I've put together a parameter file containing only the parameters needed to get the code to work with Illustris.
@Dylan, I've filled it in to the best of my ability and put it on GitHub here. Please could I clarify what is correct, and also what is currently not? In particular:
Finally, I'm uncertain regarding the parameter UnitMagneticField_in_gauss. As discussed in the PartType0 (gas) table here the units of magnetic field are given as a function of the scale factor, thus it would be different in each snapshot. @John suggests that when filling this in as a single number in the parameter file, it may be 1.0e-6. Do you agree with this?
The GravityConstantInternal and PeriodicBoundariesOn are ok.
TimeBegin 0.0078125 % z=127
TimeMax 1.0 % z=0
but I doubt these are used or matter.
I suppose the OutputList is used to know what snapshots exist? Then you would want to make this text file: each line should just be the scalefactor of the snapshot, so 100 lines in total, loading from the snapshot headers to get this value. As this is used, TimeBetSnapshot makes no sense, so could set to zero.
NumFilesWrittenInParallel doesn't matter? This is only used when running a simulation.
As for UnitMagneticField, we don't use such a parameter in AREPO. So whatever code this is based on, you should carefully check how such a parameter is used to write/convert the magnetic field values, and how this compares to the units of TNG as you link.
Thanks for the quick response. The definition of the OutputList is: "File with a list of the desired output times. Can be specified with a relative or absolute path." so I think that you are correct here. I'll give this a try. Regarding UnitMagnticField, I'll see if it's possible to just comment it out.