Switching area of interest within the domain of computational astrophysics

[mayglorytohk]

Dear All

I am a physics student that is considering strongly on persuading a PhD degree. My previous research interest and area is computational astrophysics, to be specific, astrophysical flow which is basically newtonian. Recently I was attracted towards computational and numerical relativity for which is not my original research area. How possible will it be for me to be admitted by PhD schools if I intended to switch into that area? And furthermore, is there any good recommendations of schools that do provide numerical relativity research program? I am thinking about UK/US

Thank you

[Nishikata]

I think you can apply as computational astrophysicist first to max your chances and choose any field after you got in. In US you are not bound to a group during admission, but I think you must join a group by end of first year. So there is time for switching after you get the admit.

[geekusprimus]

Speaking as someone who applied for numerical relativity, it's a fairly small field, but it's become quite a bit more popular over the last few years since the detection of gravitational waves. What's also become quite popular, though, is data analysis and instrumentation, and that has many, many more people working in it because of LIGO, and it's become quite intimately tied to numerical relativity.

Here's a list of a few schools off the top of my head with numerical relativity programs and some associated faculty who may be interesting. Some of these may not all be good choices. Saul Teukolsky, for example, is quite old and nearing retirement, but there may be other faculty at Caltech and Cornell who are doing numerical relativity.
Prestigious to Elite:

• Princeton (Pretorius)
• Caltech (Teukolsky, probably some postdocs)
• Cornell (also Teukolsky)

Good to Prestigious:

• Pennsylvania State University (Radice)
• University of Texas at Austin (Laguna, Shoemaker)

Lesser-known to Good:

• Rochester Institute of Technology (Campanelli, Zlochower)
• Brigham Young University (Neilsen, Hirschmann)
• University of New Hampshire (Foucart)
• Washington State University (Duez)

Here are some schools I know of with people who are actively involved with LIGO:
Prestigious to Elite:

• Caltech
• MIT
• Stanford

Good to Prestigious:

• University of Maryland, College Park
• Pennsylvania State University
• University of Illinois at Urbana-Champaign
• Georgia Institute of Technology

It's worth mentioning that these are just a handful of the schools with active relativity programs. Nearly every school with a relativity group has at least one person doing LIGO stuff or numerical relativity; there are only a handful of purely theoretical groups left (UCSB comes to mind).

[mayglorytohk]

Well I dunno why but it seems that physgre.com had connection problem in the past few days so sorry for the late reply
First of all thank you for the reply from Nishikata and geekusprimus
However there is one thing that I would like to ask, actually I thought that numerical relativity is a large domain
In particular, I once thought that simulation of neutron stars, black holes are to be classified as the subset of numerical relativity
Well this is quite shocking since I think what I am interested will be the applications of numerical relativity to astrophysical objects
Nevertheless, using numerical relativity to probe spacetime-matter interactions?
By the way, I saw that other than the U.S., Cambridge and southampton in the UK also provide such program
How are they different with each other? I mean, how different is PhD in UK and PhD in US is?

Best Regards

[Nishikata]

In UK, PhD does not have coursework component, so it is pretty much direct research in the lab. If you think you can learn the fundamentals yourself, or you have mastered them already, then it should be alright. Because of this, there is a shorter duration of 3 years for your PhD.

[geekusprimus]

However there is one thing that I would like to ask, actually I thought that numerical relativity is a large domain
In particular, I once thought that simulation of neutron stars, black holes are to be classified as the subset of numerical relativity
Well this is quite shocking since I think what I am interested will be the applications of numerical relativity to astrophysical objects
Nevertheless, using numerical relativity to probe spacetime-matter interactions?

GR in any flavor was not really a "respectable" field of study (at least in the words of my advisor) before the LIGO announcement in 2016. Since then, the field has grown quite a bit, but more people are interested in data analysis over numerical relativity. There are undoubtedly more schools with numerical relativity programs than what I've provided, but they're usually smaller or lesser known programs. Big schools like MIT and Harvard almost exclusively prefer experimentalists and data analysts over numerical relativists because their research output is so much higher. Other schools don't have the computational resources to support a numerical relativity program. Although production simulations requiring 10,000+ cores are typically run on computers off-site, every NR group has to have unlimited access to a smaller supercomputer for testing purposes, which means the school itself has to have one.