Custom build for data analysis in R

karmaty

Commendable
Oct 5, 2016
2
0
1,510
Hello,
I am planning on building a pc that will be mainly used to handle the computations that I can't do on my laptop. I am doing statistical analyses and machine learning mostly in R, and some in Python. I plan on installing Ubuntu on it. It will probably run for several days in a row and I need to be able to handle several Gb datasets. But I also want a reasonable price in the range of 1000-1500$. This is my first time doing this so some help would be highly appreciated. And anyone familiar with particular requirements for data analysis/science would be of a great help.
Right now, I am thinking of getting:
- Intel i7 6700K 4GHz
- 32Gb of RAM as I am currently experiencing issues working on R with large datasets. 16Gb might be enough, but I'd rather get 32Gb so I am sure that I won't have issues with super large datasets
- SSD DDR4 memory, probably 512Gb
- No dedicated GPU, at least not for now
I am not sure which motherboard to go for. A larger size ATX is not an issue.
Would the Asus SABERTOOTH Z170 MARK 1 be too much for my intended use? What about the Asus Z170-PRO ATX ? Any other one you'd recommend?
Thanks
 


That's a really good point! I think that's what I will do, as the 5820K is not much more expensive than the 6700K.
Do you have any recommendation for the X99 motherboard, or will any one with good reviews do?
Also, this is my first build, and I was wondering if there was anything to consider if I want to keep the option open of adding a graphics card later on?
 
Get a 4 DIMM matched set of RAM. Get a high quality power supply. Seasonic has always been good. You don't need a million watt power supply but you need a quality brand. Get a tier 1 or tier 2 on the power supply list -- http://www.tomshardware.com/forum/id-2547993/psu-tier-list.html

You will HAVE to get a GPU of some kind, since the 5820K doesn't have graphics built in. An NVIDIA 710, 720 or 730 would be fine.
 
Solution