I built the following rig:

5x SAPPHIRE TRI-X R9 280X 3GB GDDR5 OC (UEFI) 1x AeroCool Templarius Imperator 1150W 80 Plus Gold 1x HighPower Tech 1200w Gold Plus (HP-1200-G14C-GOLD) 1x Gigabyte GA-990FXA-UD3 motherboard 1x W.D Caviar Green 500GB SATA III HDD 1x Kingston KHX1600C9D3B1/4G 2x Unpowered risers 1x 16x 3x Powered risers 1x 16x

I have the templarius PSU powering the motherboard, the HDD and 2x GPUs with unpowered risers. The other 3 gpus are powered by the 1200w PSU and connected to the powered usb risers.

I Installed Windows 8.1 and plugged one gpu. It automatically installed the Radeon HD 7990 driver. Then connected the second gpu and followed the same process with the other gpus.

I then executed cgminer and everything looks stable at 740Khs each.

However 1h later the rig shuts down. It always shuts down aprox. 1h later..


  • What exactly do you mean by "shuts down"? The entire computer turns off, cgminer quits, or cgminer stops submitting shares, etc.?
    – Tony
    Commented Mar 9, 2014 at 22:48
  • The entire computer turns off....
    – Egidi
    Commented Mar 9, 2014 at 22:53
  • So I'm guessing, the entire computer does not turn off after 1 hour of idling if you don't mine, right? It might be that your graphics card is overheating and triggering some safety feature. What are you temps after mining for about 5 minutes?
    – Tony
    Commented Mar 9, 2014 at 23:00
  • 75º-80º celsius aprox.
    – Egidi
    Commented Mar 9, 2014 at 23:01
  • 4
    This question appears to be off-topic because it is about hardware problems with a computer you have built. You may be able to get better help on superuser.com. Commented Mar 9, 2014 at 23:36

3 Answers 3


My brothers machine was doing this also. We fixed it by pulling a card. The motherboard was pulling too much power under load.


If it's shutting off without warning, then some form of a fail safe is kicking in to keep your system intact. You shouldn't turn it back on because you're risking losing your entire setup by ignoring the signs. The problem is most likely cooling, or lack thereof. Clock down your setup if it allows, or turn down the intensity of cgminer to the one where it's not using much of your hardware's power. Then see if it cuts off at the 1 hour point. If it doesn't. Get better cooling.

Also, some version of cgminer has this weird bug where it cuts off your GPU fans after a while. If you have auto_fan enabled. Turn it off.


There are a few things I would recommend here:

  1. Make sure ALL your drivers and firmware are the most up to date available (including your bios) I had an issue with a single R9 290x that was resolved with a BIOS update.
  2. Check ALL your bios settings look for anything that might effect the power regulation to the cards.
  3. Use CGWatcher to to run the rig (not compulsory but it will make your life easier in the long run)

If this doesn't help then I suggest you try what Chris's brother did, pull a single card see if it helps. (If that resolves the issue it could be an issue with the power rail on the PCIE lanes) I realize they are powered GPU's but a friend of mine and I were once discussing large rigs such as this and he said that in some cases you need to use powered PCIE Risers like: http://www.amazon.co.uk/s?ie=UTF8&field-keywords=SuperHash&index=computers&search-type=ss

Hope this helps. I know too well the frustration of broken rigs.

Not the answer you're looking for? Browse other questions tagged or ask your own question.