Getting Started Working on the Simulations


This page is a collection of some notes I’ve put together for students interested in working on galaxy formation and simulations for the first time. Hopefully some of the background material here is helpful, and can give you a sense both of what we’re trying to do, and how we’re doing it, in more detail.


If you don’t feel like you have much background regarding the general question of galaxy formation/evolution, then I would recommend the well-written first-time introduction is in the textbook “Galaxy Formation” by Longair, which you can find in PDF form here: 

https://www.dropbox.com/s/hck3zzliddsyq4f/longair_book.pdf?dl=0

Reach chapters 3 (a good observational overview) and 19 (a good statement of some of the theoretical problems)


After reading this, you’d be in a good place to read a more technical review article on the state of the field of galaxy formation. A long but pretty comprehensive review of the field is here: 

http://arxiv.org/pdf/1006.5394v2.pdf

a somewhat shorter one is here, if you want to check that out instead:

http://arxiv.org/pdf/1207.3080v2.pdf

(there are also more recent reviews, but the ones above still capture the key issues, while some of the more recent ones are more specific, but see e.g. https://arxiv.org/abs/1612.06891)


While reading these, it may be useful to look at some lecture notes from the course I taught on these topics, specifically related to the topics our group is researching: A general lecture on the theory and simulation of galaxy formation in PDF form is here:

https://www.dropbox.com/s/c733m2fz7m59dp5/phopkins_galform.pdf?dl=0

And a discussion of many of the observed properties of galaxies, their scaling laws, and the first galaxies, is here:

https://www.dropbox.com/s/evxsigxb9adtvtd/Ay127_GalSyst.pdf?dl=0

and

https://www.dropbox.com/s/4bqiyglwvvqkaii/L11_2%20-%20UNUSED%20-%20Galform_Reion_Obs.pdf?dl=0

discussion of recent updates from our work on the FIRE project is here (this is from a conference talk, so its more high-level, youre not expected to follow it all from pdfs alone but it provides an overview) 

https://www.dropbox.com/s/xh86sv5blwkyghe/phopkins_FIRE_updates.pdf?dl=0

and discussion of star formation on many different scales is here:

https://www.dropbox.com/s/cz8kpqyb6fpxfw0/phopkins_sf_sexten_talk.pdf?dl=0

If you want to actually see one of my talks about galaxy formation, the best one for "introduction level" can be downloaded here:

https://www.dropbox.com/s/ckwjxc8ch5a33x7/aas_229_331_01_hopkins.mp4?dl=0


Thats a pretty good starting point in terms of background in the field. Now, to get more into the simulations themselves, we’re going to need to get into more technical papers. I'd definitely suggest you check out our most recent paper on the simulation results. Its going to perhaps jump over some material (maybe not the best first introduction), but its what we're actually working on and it pretty well states what the cutting edge of the field is right now. Its at 

http://arxiv.org/abs/1311.2073

(You can actually see a talk I gave on this recently online at http://online.kitp.ucsb.edu/online/stars-c14/hopkins/  . Also, a bigger, arguably better, certainly longer, and more technically-focused version of this paper is here: https://arxiv.org/abs/1702.06148)


Once you read this, if you want to start learning the code, how to read in snapshots, and actually working with the data, you should go to the GIZMO section of this website:

www.tapir.caltech.edu/~phopkins/Site/GIZMO.html

and check out the GIZMO (our code) user’s guide. 

http://www.tapir.caltech.edu/~phopkins/Site/GIZMO_files/gizmo_documentation.html

It will walk you through how to do things like read in a snapshot, do some basic data manipulation, etc (read the section on “snapshots and initial conditions files” in particular). If you want to run the simulations, it’s got everything you need; test problem initial conditions and parameterfiles and a walkthrough of how to do it. 


If you need an example snapshot of one of the simulations to work with, you can grab one here: 

https://dl.dropboxusercontent.com/u/16659252/snapshot_440.hdf5

It’s a snapshot of a cosmological simulation of a star-forming galaxy, run to z=0 (the time of the snapshot). Play with it - you should be able to find the main galaxy in the simulation and “look at it” with your preferred set of tools (I and most of us are using Python for this kind of analysis). 


You should make sure I add you to both the bitbucket repository for GIZMO, so you can grab a current copy of the code, and to the FIRE Wiki. That’s the next place to check out, here:

https://sites.google.com/site/gadgetpzooms/home

This is a wiki page specifically devoted to our collaborations work on the simulations of galaxy formation. Lots of good stuff in there including discussions of ongoing simulations, lists of their locations (where to find them), paper topics, instructions for reading in the snapshots from these specific runs (and what the results mean), plots of various quanities from simulations we have run and projects in progress, and so on. If you can follow what’s in here, you’re definitely more than ready to do detailed analysis yourself.



If you're interested in using our simulations, and want to directly access more of the data than is easily here, great! We have a huge archive with all of our old simulations. The tricky part is in getting access. The simulations are all archived an a national system -- the "ranch" archival system at XSEDE. If you haven't used the system before, there is a user guide here: 

https://www.tacc.utexas.edu/user-services/user-guides/ranch-user-guide

You will need an XSEDE or TACC account. If you have one, just let me know, I will add you to the project, and you can immediately access/download any of the simulations! They are all archived in the following directory: 

/home3/01799/phopkins/

which you should have full read-access to as soon as I add your username to the project 'allowed' list. 

If you do not already have an XSEDE account, you need to get one, which you can do at their user site: 

https://portal.xsede.org/

(see "create account" under the user login. make sure your account is also a TACC account, since they are the sub-group that administers ranch). 

The files are huge, and there are hundreds per simulation, so there isn't any way to keep them all locally -- hence the massive archival system. 




© Philip Hopkins 2015