Author

Topic: [MINING rig] [Windows 10] NMI_HARDWARE_FAILURE when multiple gpus plugged in (Read 153 times)

newbie
Activity: 4
Merit: 0
A61TR1
06147-3
Bios spi1 ver.0122.

PS: I tested the rig using my PC’s mobo (gigabyte ga-somerhing) and all 5 gpus are working.
full member
Activity: 585
Merit: 110
what is your motherboard model number?
newbie
Activity: 4
Merit: 0
Hello guys,

I just setup my mining rig having only one plugged gpu (AMD r9 380 Nitro). All good, successfully installed windows 10, windows 10 updates, crimson amd 15.12 drivers, claymore 9.7 beta.

I restarted the rig after all the programs were installed. I mined a bit using claymore to monitor the temperatures, power consumption , stability, etc. Again, everything worked fine.

I shut down the rig to plug the rest of the gpus (another 4 AMD r9 380).

After that I started the rig, but the windows won't start due to blue screen NMI_HARDWARE_FAILURE.

Ok, then I unplugged 3 of the cards resulting in only 2 plugged-in gpu cards. I restarted the rig several times to see if something changes in the behaviour. I observed two cases :
  • blue screen NMI_HARDWARE_FAILURE
  • enters windows, but cannot do anything. I just can move the mouse(cursor). I cannot open anything, Neither CTRL+ALT+DEL is working

Tried to change the pci slots, the risers, etc. The problems remain.

I unplugged one more graphic card -> the rig is working fine using only one card.

The components of the rig are:
  • HP SERVER ProLiant ML310 G5p MoBo
  • 450W PSU + 1200W Modular PSU (synchronized)
  • 4 GB ram
  • Intel Dual Core 3.0 GHz
  • 5 x AMD R9 380 2/4GB
  • SSD 120Gb

I went through bios but those are the only configurations for PCI slots (when 4 gpus are plugged in):
https://image.ibb.co/cDk21H/IMG_0568.jpg

The exactly same behaviour I encountered with default AMD drivers installed by windows

Any ideas?
Thanks!
Jump to: