Computational Astro - Go into 5th year to get Comp Sci deg?

  • This has become our largest and most active forum because the physics GRE is just one aspect of getting accepted into a graduate physics program.
  • There are applications, personal statements, letters of recommendation, visiting schools, anxiety of waiting for acceptances, deciding between schools, finding out where others are going, etc.

Post Reply
GeorgeThe4th
Posts: 22
Joined: Thu Feb 10, 2011 8:44 pm

Computational Astro - Go into 5th year to get Comp Sci deg?

Post by GeorgeThe4th » Thu Feb 10, 2011 9:09 pm

EDIT:

Thanks for the responses, they gave me some much appreciated tips and advice. I figured I would give an update:

I talked to my computational astro professor last week about possibly doing research this summer. He said it definitely is a possibility and that he would think about a possible project that would be suitable. Meanwhile I asked him his opinion on getting more of a computer science background. He told me to not worry about it, and that anything I would need to know could probably be picked up by tutorials from a supercomputing group that provides help on our campus.

By the way, I will be working on numerical magnetohydrodynamics with him this summer, and am going to have to learn Fortran and some basic plasma physics in the mean time.


ORIGINAL POST:
_____________________________________

Hi, I am a junior Astrophysics major at the University of Minnesota. I have put a lot of thought into graduate school and the entire application process. I actually just took the General GRE last month, so I consider that the first official step. I am torn between a few different options for how I finish my last year and improve my chances of getting into a good program.

This semester I am taking a class on computational methods in the physical sciences, and it has inspired me to pursue computational astrophysics. I actually only need my senior thesis (which I am doing this summer), and otherwise all that is left are 15 elective credits. So my senior year is looking to be unusually uneventful as far as course load is concerned. This is basically a result of me taking a heavy load my first 3 years because I thought I was going to double major astrophysics and physics, but have recently been questioning that option.

One of my big questions is whether a strong foundation in computer science could be beneficial for me, if I want to dive into computational astrophysics. All I have taken in that regime is an introductory C++ course last fall. My question is whether it would be beneficial for me to complete a 5th year and get a computer science degree to supplement my astro degree (I figured out it could be achieved if I take a heavy load).

It basically is a question as to whether delaying graduate school by one year will help me more than it will hurt me. So far I have no research experience whatsoever, and my senior thesis this summer with my computational professor will be my first exposure to that. If I do stick to the 4 year plan, then I am planning to cram two more separate research projects into this fall with two other professors, so that I can have three letters of recommendation from people who all have observed my qualities on a closer level than in the classroom.

If I went into a fifth year however, I would not have to cram those 3 research experiences into a summer and a fall. I would have more luxury to spread them out. Or is it reasonable for one of my letters to come from pure classroom experience? I am taking a graduate level observational methods class right now. Could that professor suitably vouch for me from my performance in a graduate class alone?

Sorry for all the randomly scattered questions. The main point of this topic though is the dichotomy between 4 years and 5 years in amassing the credentials that will help me with my future. In a nutshell: do I double major in computer science, or stick to just an astrophysics degree and assume that my computational abilities will be developed through use in graduate school?
Last edited by GeorgeThe4th on Wed Feb 23, 2011 8:32 pm, edited 2 times in total.

astrok
Posts: 21
Joined: Fri Feb 04, 2011 10:36 am

Re: Computational Astro - Go into 5th year to get Comp Sci deg?

Post by astrok » Thu Feb 10, 2011 9:52 pm

The lack of research will be the thing that hurts you the most in grad school applications.

As far as the extra coursework, it could be beneficial, but I recommend taking extra courses in both math and computer science that appear useful rather than going for the computer science degree. I got a CS minor, and I can tell you that you will never care about a lot of the material in CS classes. As far as programming languages, for the most part, once you've learned one, it's not too tough to pick up another. Upper level CS classes will tell you more about memory allocation, objects, and the best way to sort an array than you will ever need to know for computational astrophysics instead of teaching you new programming languages (and if they do teach you a new language my bets are on java, which is SLOW i.e. bad for computational astrophysics). Things that will be beneficial are classes like numerical methods (probably offered by your math department), and parallel computing.

My suggestion is that you talk to the professor you will be doing research with this summer, and ask for an accurate assessment of your current preparation for graduate school and courses that may benefit you if you end up in computational astrophysics. Moreover, I would not attempt to cram two research experiences into one semester, particularly one where you need to be working on graduate school applications. If you put too much on your plate you could very well end up with two mediocre references. Having one letter from someone who only knows you on the classroom level is not necessarily bad, as long as they know you well and can say more than "he got an A in my class".

User avatar
grae313
Posts: 2296
Joined: Tue May 29, 2007 8:46 pm

Re: Computational Astro - Go into 5th year to get Comp Sci deg?

Post by grae313 » Fri Feb 11, 2011 1:50 am

This is great advice, and to add to it, focus on research quality, not quantity.

m42
Posts: 11
Joined: Wed Feb 09, 2011 3:28 pm

Re: Computational Astro - Go into 5th year to get Comp Sci deg?

Post by m42 » Fri Feb 11, 2011 6:04 pm

I decided to veer off and tried doing a computer science degree and an astronomy degree my sophomore year, and I do regret the stress and lack of sleep it cost me. I eventually did settle for just an astronomy degree.

That being said, one of the computer sci courses I took was a 2 semester course that was programming for Computer Science Majors, not the regular C++, or Fortran Intro classes. It was much more beneficial, as in the latter intro classes they were fairly basic and consisted of weeks of simplistic scripts, i.e. Hello World. The Programming for CS Majors was more intense, and more beneficial in the long run - even though it was in Java. Once you pick up a programming language, the general way to setup complex programs, and a little info on time/memory allocation issues you're good to go in any language.

So if UM has a similar type of intro course specifically for CS majors, that would be most beneficial. Instead of going for the major, it would be better served to devote your remaining time into research, and studying for the physics gre well in advance.

And Go Gophers! - I'm a former Minnesotan 8)

GeorgeThe4th
Posts: 22
Joined: Thu Feb 10, 2011 8:44 pm

Re: Computational Astro - Go into 5th year to get Comp Sci deg?

Post by GeorgeThe4th » Wed Feb 23, 2011 8:30 pm

Thanks for the responses, they gave me some much appreciated tips and advice. I figured I would give an update:

I talked to my computational astro professor last week about possibly doing research this summer. He said it definitely is a possibility and that he would think about a possible project that would be suitable. Meanwhile I asked him his opinion on getting more of a computer science background. He told me to not worry about it, and that anything I would need to know could probably be picked up by tutorials from a supercomputing group that provides help on our campus.

By the way, I will be working on numerical magnetohydrodynamics with him this summer, and am going to have to learn Fortran and some basic plasma physics in the mean time.



Post Reply