Table of Contents

  1. Why I Built It
  2. Hardware
  3. Software
  4. Screenshots

Build A Computer for Deep Learning

Posted on 19 Jun 2016, tagged life

These days, I bought some hardware and build a computer by myself. It’s fun and I’ve learned lots of things from it. Here is a note to record it.

Why I Built It

This computer is used for working and learning. I build it for two reasons:

  • Use Linux. I’ve used Linux for many years and then used MacOS X. I missed the days of using Linux, which I can control everything.
  • Run deep learning programs.


CPU Intel i5-6600k It is easy to overclock and has built in graphics to support 4k at 60Hz.
RAM G.Skill DDR4 2400Hz * 2
SSD Samsung 850 EVO 120G SATA3 I use this for the system volume.
Disk WD blue disk 2TB * 2 I use this for the data volume. These two disks are built with Raid 1 so the data is safe. I've considered to buy three 1TB disks and built Raid 5, but the price is not so good.
Motherboard GIGABYTE Z160X-UD3 This motherboard is extendable. However, it doesn't support display port for 4k at 60Hz. (I found this after I bought it and I couldn't return it ...)
Power GreatWall 600W GW-7000D It has enough power.
Case SAMA Tank This case is big enough. However, it has just two 3.5 disk positions and three 2.5 disk positions. But I can deal with it.
CPU Cooler DeepCool This cooler is by wind and the noise is small.
Monitor DELL P2715Q This monitor has 4k support. The display effect is very fantasy.
Graphics NVIDIA GTX 1080 I haven't buy it yet. Use this to run deep learning programs.


I may share all my configurations some day. Here is a brief list:

OS Archlinux I've used ArchLinux for many years and I love the philosophies like rolling update, just install what you need ans so on.
Window Manager i3wm Tilling window manger is very awesome.
dmenu Rofi It supports fuzzy search and searching opening windows.
Monitor Conky


Split with Chrome and Terminal: screenshot1

Conky and Rofi: screenshot2

Read a PDF: screenshot3

comments powered by Disqus