VEGA 64 + Ryzen 2700X - System shutting down 10-15 mins into gaming.

[sosij93]

Hi All,

I’m having an issue at the moment with a new pc I built over the weekend: 15 minutes into a game my pc shuts down as if there was a power cut, no warning/bsod/etc. At first I thought it must be a temp issue or the psu isn’t up for it, but after running an aida stress test for CPU, GPU, Storage and Memory all at once the system held up without any issues (stopped the test at around 90 mins out of boredom). During this test CPU temps maxed at 80C and averaged just under that, GPU held up fine and looking at power there was never any noticeable deviation from the voltages to suggest a psu issue either. This has left me a bit confused and am unsure what is causing these shutdowns – I expected the stress test to shut my pc down like gaming did, especially considering the stress test put the whole pc under a heavier load than gaming does.

Any ideas or other avenues I should explore? – appreciate any help here,  This is my 1st time posting on such forums so please ask for more info if I’ve missed anything. Specs:

AMD Ryzen 7 2700x @ Stock speed (for now) – Brand new
Corsair H100i v2 AIO liquid cooler – Brand new
Sapphire nitro+ RX Vega 64 (not the limited edition one) – had working in old system no issues.
MSI X470 Gaming Pro Carbon – Brand new
16GB Corsair Vengeance 3000mhz ram (2x8gb) – Brand new
Thermaltake Toughpower Grand 850w 80+Gold – Brand new
1TB HDD – had working in old system
2 x 240 SSD – OS Drive was in use in old system
Corsair 570x Case – Brand new

 

[profoundnoah]

I myself have experienced similar issues with my 2700X, ended up having to RMA it. You probably won't need to, though. Can you get a list of temps when you're running that game (using software such as HWinfo etc.)? My CPU ended up running at 96C when streaming/recording, but wouldn't budge above 85 when doing AIDA. Always good to double-check.

 

[mdd1963]

Try opening the case, and blowing a small case fan across assorted mainboard power circuitry...

Not familiar with all the X370/470 BIOS options, but, I'd be searching on Google for any potential power limiting/thermal features, especially that might pertain to supplying power thru the PCI-e slot, as the Vega 64 is rumored to draw up to 300-350 watts under full load, and ~75 of those watts will be thru the PCI-e slot...
(Underclocking the Vega64, although not an attractive option, might spare you 30-80 watts, and allow more headroom if a power constraint....)
You might also experiment with using an UPS of sufficient wattage, just to rule out the classic 'rig powers off when home's AC turns on'-syndrome...; if any rig might be vulnerable/more prone to it, it's a Vega 64

 

[sosij93]

I myself have experienced similar issues with my 2700X, ended up having to RMA it. You probably won't need to, though. Can you get a list of temps when you're running that game (using software such as HWinfo etc.)? My CPU ended up running at 96C when streaming/recording, but wouldn't budge above 85 when doing AIDA. Always good to double-check.

Try opening the case, and blowing a small case fan across assorted mainboard power circuitry...

Not familiar with all the X370/470 BIOS options, but, I'd be searching on Google for any potential power limiting/thermal features, especially that might pertain to supplying power thru the PCI-e slot, as the Vega 64 is rumored to draw up to 300-350 watts under full load, and ~75 of those watts will be thru the PCI-e slot...
(Underclocking the Vega64, although not an attractive option, might spare you 30-80 watts, and allow more headroom if a power constraint....)
You might also experiment with using an UPS of sufficient wattage, just to rule out the classic 'rig powers off when home's AC turns on'-syndrome...; if any rig might be vulnerable/more prone to it, it's a Vega 64

Thanks both mdd and profound for your help on this one.

Profound, I've used cpuid hwmonitor, aida's cpu temp sensor, corsair's icue/link all at once to monitor cpu temps during gaming as well as amd's global wattman for gpu monitoring - this is where the confusion was as cpu and gpu temps were as expected during gaming (50-55 for cpu and 70-75 for gpu).

Mdd, I poked around the bios and Google but couldn't find anything useful but your point helped me in the direction of what I think is now the solution. I've isolated one temp sensor (called tmpin7 in cpuid) which is consistently higher than the others, even higher than the cputin, and believe it refers to the part of mobo sat right next to the vega. So although the gpu was never breaking 75 it was heating that part of the mobo to 85C+ (likely triggering a system shutdown) as the sheer size of the card completely smothers that part of the mobo. I've since decreased gpu target temp to 65 with the tradeoff that the fans will be rinsed much more to achieve that target as well as configured the case fans to be based on that particular mobo temp sensor to ramp up when that part gets hot and hopefully stop it overheating. Not had the problem occur again since but will monitor carefully and post on here again if the issue comes back.

Appreciate your help, tthank you