Last night my PC shut down a couple times unexpectedly whilst under heavy load after ~1 hour of gaming each time. I was hoping someone could help me identify the cause of the shutdown, or reccommend some steps to diagnose the problem.
As I mentioned, my PC shut down under heavy load either by the PSU or motherboard cutting power (as it would if you were running an unstable overclock to prevent damage).
I recently swapped out several components. My system specs below:
CPU: I7-4770k (New, was overclocked 0.4GHz but reverted back to stock and problem persists)
Mobo: Maximus VII Ranger
GPU: Sapphire R9 290 reference (newest addition, fault occured after upgrading from 2x HD 6850)
16Gb kingston 2400 RAM
2 hard drives, optical drive, 3 case fans
750W PSU (Can't remember the brand but it's not a cheap shitty one)
Logic dictates that the GPU being newest is the cause - and I assume it is either overloading the PSU, or overheating. I intend to more closely monitor the gpu and vram temps and fan speeds tonight, but my first thought was I need a bigger PSU (Sapphire reccommend 750W with the 290 so I'm on the limit)
I'd appreciate any thoughts on what the most likely cause is, and any software you can reccommend for diagnosing the fault. Ideally a program that can log temperatures up to possible shutdowns. I unfortunately don't have another PSU to swap in, does anyone know an easy way to test PSU load?
If you've read this far, then thanks for your time and I appreciate any help or feedback from the community
I suspect gpu is overheating. Try using gpu temp or similar software to monitor your gpu temps while gaming. This card can get very hot. Especially with stock amd cooling.
After some more extensive monitoring it looks like was indeed the gpu temp. The default max fan speed just isn't high enough to keep it cool so I had to ramp it up, despite it being at a standard clock rate. Kinda stupid really, I knew they ran hot but I figured they'd at least increase the fan rate to stay below overheating, even if it meant extra noise. Apparently I was wrong!
Thanks for the help anyway, I ended up using GPU-Z for my monitoring and it worked wonders