Workstation build

fizbo4

Junior Member
Nov 27, 2015
3
0
0
1. Workstation, some scientific computing (with GPU), maybe light gaming
2. 1600ish but it's flexible
3. US
4.
5. nVidia graphics
6. already have monitors
7. not planning on overclocking
8. probably dual 1920x1080
9. ~now
10. will run linux

Case: Corsair Carbide Series Air 240 High Airflow MicroATX and Mini-ITX PC Case
Motherboard: ASUS Z170M-PLUS Micro ATX
Processor: Intel i7-6700K
RAM: Corsair Vengeance LPX 32GB (4 x 8GB) DDR4 DRAM 2400MHz
Storage: Samsung 850 EVO 500GB 2.5-Inch SATA III Internal SSD
PSU: Corsair CX600 CP-9020048-US 600W
Graphics: EVGA GeForce GTX 980 Ti

$1,633.84 pre-tax on Amazon today

What did I forget?
 

fizbo4

Junior Member
Nov 27, 2015
3
0
0
Couple questions:
- could I get away with a smaller case? I feel like this one will be pretty empty, though I have the full size graphics card.
- could I get the SSD in m2 form factor? I'm not sure if it would use up PCIe lanes, when I'd like all 16 of them available for the graphics card.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,232
3,812
75
What did I forget?
CPU cooler? That would fill up some of your case. But, yes, you could get a micro-ATX case and mobo.

You could get the Samsung 950 in M.2. It would pretty much blow your budget, though. :\

Can you elaborate on "scientific computing"? An i7-5820K or 5930K (and associated mobo) might make more sense. For some - but not most - applications an R9 280X is the fastest GPU you can get.
 

fizbo4

Junior Member
Nov 27, 2015
3
0
0
Thanks for the reply!

CPU cooler? That would fill up some of your case. But, yes, you could get a micro-ATX case and mobo.
Ah indeed, I'll add a Cooler Master Hyper 212 EVO to the list.

You could get the Samsung 950 in M.2. It would pretty much blow your budget, though. :\

You mean just as an upgrade from the 840 EVO? or does the 950 get around the PCIe lanes issue?

Can you elaborate on "scientific computing"? An i7-5820K or 5930K (and associated mobo) might make more sense. For some - but not most - applications an R9 280X is the fastest GPU you can get.

Training deep neural nets mostly. For that, the 980ti actually comes pretty highly recommended, I think.
 
Last edited: