Laptops for Ph.D. programs

[BL250604]

Hi all-

I need to buy a new laptop (currently a mac user- but will be switching) and wanted to see if people had any advice on what to look for when purchasing a laptop for grad school. I will for sure make sure that I have 16GB+ of RAM and an i7. Other than that, are there any things (from your experience) that have been really useful to have in your laptop? I would think that windows would be good when working with R Studio and other programming languages as well. Thanks!

 

[bayessays]

Is there a reason you're switching from Mac? I think they're the best of both worlds for a statistics grad student. Depending on your research and how much programming you do, having a Unix computer would probably be easier for most people than Windows. Other than that, yes, I'd make sure to have a lot of RAM because I've always had issues with RStudio crashing.

[BL250604]

9 minutes ago, bayessays said:

Is there a reason you're switching from Mac? I think they're the best of both worlds for a statistics grad student. Depending on your research and how much programming you do, having a Unix computer would probably be easier for most people than Windows. Other than that, yes, I'd make sure to have a lot of RAM because I've always had issues with RStudio crashing.

Hi! Actually, I've been doing some research using INLA this semester and have had issues with the package developers not updating the package to be compatible with new versions of R studio. This has happened before with spatial packages ive used as well, and at times its been frustrating for workarounds. Do you think a mac is okay, and that this is probably a rarer thing? The R studio stuff is my main concern, as I do love my mac otherwise.

[bayessays]

5 minutes ago, BL250604 said:

Hi! Actually, I've been doing some research using INLA this semester and have had issues with the package developers not updating the package to be compatible with new versions of R studio. This has happened before with spatial packages ive used as well, and at times its been frustrating for workarounds. Do you think a mac is okay, and that this is probably a rarer thing? The R studio stuff is my main concern, as I do love my mac otherwise.

I'm not super familiar with this exact problem - is it specific to Macs? I've had issues with RStudio on other operating systems, too, mostly related to crashing with memory overloads. Another was some package compatibility issue with Linux. I don't think there's any reason to believe RStudio would work better on Windows. I know people who have used all 3 major operating systems without issue, so my advice would be to use the one you're most comfortable with.  If you go to a biostatistics department that uses SAS, though, you might have some hassles if you use Linux where it's not available.

[orchidnora]

I think you should also post your budget, keeping in mind that some departments/advisors might provide you with a work computer or funding for one as well. 

One thing that's nice about Mac is that the hyperthreading can be easily exploited in R through the "parallel" package. Using the "parallel" package, you can easily speed up certain types/sections of code drastically. This is convenient if you plan on running your code locally. 

If you get into heavily computational work, then you will be probably be mostly working on HPC. In that case, the quality of your computer doesn't matter as much (a $500 laptop would suffice), as long as you can connect remotely to the HPC server. Connecting to the server is obviously easier through Unix, but there are programs you can download to access it through Windows too. 

Personally, I like working locally in RStudio as well as using HPC, so I did recently buy a nice ASUS laptop with Windows. I would have liked a Mac, but a Mac with similar specifications is far outside my budget.

Edited February 16, 2019 by orchidnora

[TITX]

If you're in the market for a desktop have you ever considered building your own? I was able to customize mine and am really happy with the result. There are two sites that really helped me: the reddit build your own pc forum and pcpartpicker. Pcpartpicker is really helpful because it will let you know if certain parts that you like are compatible with each other. 

https://www.reddit.com/r/buildapc

https://pcpartpicker.com/

[BL250604]

Thanks for the advice everyone! @bayessaysI will definitely follow your advice and go with what I'm more comfortable with. Maybe I won't switch and get another mac. Thanks all!

 

B