PowerSpec G465 GPU Overheating

Alpha087

Roughly a week ago I picked up a PowerSpec G465 and out of the box I'm noticing that the RTX 2080 Super that it came with is reaching up to 87C while under full load.  I tried contacting Micro Center's live chat support about it and whomever I spoke to said that the average temp under load for one of these cards should be around 75C and that the temps I was reaching wasn't very far off...  Which is hilariously wrong.  And further research yielded that the card's thermal throttling limit is 88C.

Anyways, I've tried everything from setting a temp limit in MSI Afterburner, which the card will not really seem to obey, to setting the fans to 100% speed, which will still not get the card below 80C.  The only temporary workaround I've found is to set a hard core clockspeed and voltage limit, which works, though I don't think I should be needing to do that on a brand new piece of hardware. 

So my question then is, what course of action do I take?  Do I return the entire build to Micro Center?  Or is this "normal" operating temperature as the service rep I spoke to in live chat claimed?  I worry that if I let this card operate at those temps it will burn itself out after not very long and/or potentially damage other hardware.

PowerSpec_MichaelB

Welcome to the forum, @Alpha087. There may have been a misunderstanding with the previous technical support rep in regards to what is "normal" and what is "safe". As you've mentioned, the card begins to thermal throttle around 88C as that is the threshold where the silicon starts to degrade under prolonged usage. As to why you cannot tame the thermals despite imposing a temperature limit, you're likely running into what is known as "thermal runaway". Your card is getting hotter, causing it to draw more power, the more power it draws, the hotter it gets, rinse and repeat. Until the reason behind why the card is getting excessively hotter is resolved, you won't be able to tame the thermals with a simple thermal limit.

Your clock/voltage limit will do the trick as it tames the amount of power the card can pull which effectively limits the potential of thermal runaway, but doing so comes at a pretty hefty cost in performance, and this is far from ideal. 

I didn't have a G465 on hand to test with, but I had the exact parts on hand to put one together as a little mockup. To give you a better understanding of how your system should be performing, Under a combined load of Furmark (4k resolution, 8x MSAA) and Prime95 (48k FFT length min/max, 24576mb RAM) I was unable to exceed 78C on the GPU with fans on auto. The GPU maintained a turbo clock of 1935mhz. The ambient temperature of the room I am testing in is 75F and I let the test run for roughly 30 minutes.

It's worth noting that this is a completely unrealistic synthetic workload, so real-world usage, you probably shouldn't be reaching this temperature unless you were overclocking, or running for hours and hours on-end with no break and your ambient room temperature was pretty high. I'd say with these results in-mind, 87C is not normal unless your room temperature is 82F or higher.

As for what you should do next, I would recommend bringing it back in for an exchange. If it's not convenient to bring it in right now, then you are fine to maintain your underclock/power limitation as that will not harm the hardware in any capacity. Even letting it throttle itself at 88C won't technically damage the hardware, it will just cause massive dips in performance. As soon as it is convenient for you to bring it to your local Micro Center store, bring it in and our team will assist with getting it exchanged for you. 

If you have any questions in the meantime, do not hesitate to let us know!

Alpha087

So just an update on this;  I brought the PC in to Micro Center and due to supply shortages the best they could do was check it in for repair/service.  After about a week, I was contacted by a tech from Micro Center who claimed that after running benchmarks for an hour or two, he couldn't replicate the temps I was able to reach while gaming.  I tried to walk him through which games he might be able to replicate it in, though he ultimately, for whatever reasons, could not replicate the overheating problems and advised me to take a video recording of the computer overheating myself.  To top it off, while I was picking up my PC, I spoke with someone at the tech support desk, claiming to be a district manager, telling me that according to the lab techs who I assume plan out the specs for these machines that these cards "are rated to run up to 100 degrees", which...  Obviously can not correct if he was talking temperatures in Celsius.  

Anyways, I'm kind of at a loss for how to proceed now.  I have the PC back and I've since been able to test several other games, more and more of which are seeming to hit around 86C with relative ease.  Should I go in asking for a manager next time?  Seems like the people they have at the front desk, at least when I last visited, did not actually know what they were talking about.

PowerSpec_MichaelB

Alpha087 said:

So just an update on this;  I brought the PC in to Micro Center and due to supply shortages the best they could do was check it in for repair/service.  After about a week, I was contacted by a tech from Micro Center who claimed that after running benchmarks for an hour or two, he couldn't replicate the temps I was able to reach while gaming.  I tried to walk him through which games he might be able to replicate it in, though he ultimately, for whatever reasons, could not replicate the overheating problems and advised me to take a video recording of the computer overheating myself.  To top it off, while I was picking up my PC, I spoke with someone at the tech support desk, claiming to be a district manager, telling me that according to the lab techs who I assume plan out the specs for these machines that these cards "are rated to run up to 100 degrees", which...  Obviously can not correct if he was talking temperatures in Celsius.  

Anyways, I'm kind of at a loss for how to proceed now.  I have the PC back and I've since been able to test several other games, more and more of which are seeming to hit around 86C with relative ease.  Should I go in asking for a manager next time?  Seems like the people they have at the front desk, at least when I last visited, did not actually know what they were talking about.

I am sorry to hear that they were unable to replicate the issue in their testing. If you'd like, I can reach out via private message and work to setup a remote connection to your system so that you can load your typical games and we can take a look at the clock speeds/voltages under test to see what occurs. 

To help figure this out, can you provide some additional information about your setup at home? Is the system located in a position that it's able to get adequate information from the front intake fans? If the desk is stuffed into a small cubby area, make sure that the perforated holes on the front plate are far enough from side panels that they can still intake some air. What is the ambient temperature in your house? Typically anywhere from 70F-80F is fine, but much higher than that and you'll start to see your hardware's temperature creep up as a result.

I'll contact you shortly to see what we can do and hopefully get this taken care of for you.