Create an account on the HP Community to personalize your profile and ask a question
02-23-2022 02:19 PM - last edited on 03-21-2022 08:27 AM by JessikaV
About a year ago I bought a new Omen Desktop with the 3080. Immediately I found the cooling to be underwelming, for both the CPU and GPU, but I reckoned it would be more than sufficient given HP and Cooler Master figured it to be so. I regularly checked the Omen hub, CPU-z and HWmonitor to see what's happening when it's ramping up.
The 3080 seemed to have have a relatively high memory temperature of 106°C (stable), with the core and hot-spot being +-80°C and +-92°C respectively.
The card died on me a few weeks ago, and monday it was swapped out for a new HP 3080 unit. This one however has the hot-spot running at 108°C (stable) with the core and memory being 81°C and 86°C respectively.
Could these temps have anything to do with the premature failure or are these normal?
-For context: the PC is placed in a well ventilated area, on a solid surface with a constant ambient temperature of +-21°C. All results are without any overclocking, straight out of the box running games like: DCS, War Thunder, GTA, RDR and Hitman as well as using it with the Reverb G2.-
Thanks for any assistance!
02-23-2022 06:10 PM - edited 02-23-2022 06:19 PM
No, these temps are NOT normal and would definitely contribute -actually: in my opinion would be the primary cause of premature failure of your GPU.
I also have an RTX 3080 -the Asus GeForce RTX 3080 12GB-OC Edition to be precise. Even though I also game seriously, I have not seen anything close to the temps you observed on your 3080. One difference, in my particular setup, is that my 3080 happens to be parked outside a souped-up HP EliteDesk 800 G3 SFF, which helps a lot keeping GPU temps down.
For comparison's sake, why don't you run, for example, the relative benign "(GL) msi-01" stress test on MSI Kombuster (version 18.104.22.168) at 1920x1080 resolution for ten minutes, like this:
This is what I got:
Notice that my 3080's "hotspot" was only 71.6° C and a Memory temp of 82.0° C.
02-24-2022 11:45 AM
Thanks for the quick and elaborate response!
I figured the temps were bad, but I looked the other way seeing how it was a pre-built with a warranty and all. The GPU got a new lease of life (and warranty) but that doesn't mean I'm looking to cook this one as well ofcourse.
My 3080 is an HP version of the card, which is a reference card that looks quite similar to many other models in the general sense. (triple fan configuration)
(Not my picture, same card)
I ran the tests exactly the way you ran them for some 10 minutes.
This is actually the first time I ran a test that show the cap, which is the temperature from start to finish.
I've been thinking about getting a bigger case with better cooling, for the CPU as well from the start. The 10700k is cooled by a single 120mm using the hot air exhausted by the GPU and will hover around the 95°C while gaming (no overclocking) with the built in turbo boost turned off, for which you have to go a long way into the system.
I've been running it like this since I got it because it ran the games alright and I didn't feel like dropping another few hundred bucks on things that weren't necessary, but this has me thinking again...
Again, thanks for the help and confirmation so far, it's much appreciated!
02-24-2022 12:15 PM - edited 02-24-2022 01:31 PM
I did some more testing to see what's what.
Out of the box:
The GPU sat at 77°C, the hotspot at 104°C and the memory at 82°C. The CPU sits at 67°C.
Without the sidepanel:
The GPU sat at 75°C, the hotspot at 102°C and the memory at 78°C. The CPU sits at 59°C.
Without the sidepanel and with the fans locked to 100% in Afterburner.
The GPU sat at 72°C, the hotspot at 98°C and the memory at 76°C. The CPU sits at 60°C.
The temperatures drop quite a bit, especially the CPU and the hotspot. To be honest, the core and memory temperatures don't look half bad, but the hotspot is still on fire.
We also see it going from being capped by the thermals to being capped by the power draw with the drop in temperature.
02-25-2022 12:17 AM
If you can get a hold of an Omen 45L case as a standalone purchase and transfer your PC components into that chassis, that would be one fantabulous option to consider: HP's Omen 45L case puts the radiator on the outside | TechSpot
The 45L case offers a massive internal cooling environment improvement.
Another option to consider is to think outside the box -yes, literally, in this "case": add more cooling fans. As a matter of fact, I did just that today on my most recent Legacy HP Upgrade project. And it actually worked out A-OK: Solved: Upgrading HP EliteDesk 800 G3 SFF - Page 7 - HP Support Community - 8251218
The thinking was to substantially increase the positive internal air pressure and the downward airflow through my CPU heatsink. And that I accomplished: