I just swapped my 2080FE for a 3080ti (Asus TUF).
My 2080 started to go bad after some play time, so I suspected temperature issues. It was clean etc but maybe old thermal paste. I do a lot of flightsim development so needed a new card right away. I think it died an early death as I have spent countless hours in the MSFS menus using the SDK tools, with the menu pinning my gpu at 100% the whole time. Thank goodness for low power mode in SU10
This is false. While hot, these cards are designed to run at this TDP.
It’s micro electronics on a tiny scale so mostly there is very little soldering that can be done by hand and very few places that could competitively do such a repair outside of Asia. I’m pretty sure Gigabyte will just bin it although I guess they might recycle the gpu chip.
Cards get broke when they are hot and not when they are cold.
If there is one crusial part that’s weak or not soldered properly and it gets too hot, it’s bye, bye. The temp. measurement is not at component level only at chip level, so you can’t see what happens until it’s to late.
This is a very elementary way of thinking about it, and isn’t limited to upper-level cards as suggested. If you have a bad sodder on a 3060, you’ll find out quickly.
Running cards hotter shortens their lifespan but you can be pretty sure they’ll last through the warranty period … unless you’re mining but that will possibly void it anyway.
Well I’ve bitten the bullet. Can’t wait 10 days for my refund and be without a gpu, so I’ve sold my kidney and bought one of these 
MSI Suprim X 3090ti
Believe it or not, but I should be £200 better off with the refund and the cost of this card, as i well over paid for the RTX 3080ti last year.
Hopefully this should see me through for a couple years.
Now to tell the wife 
I got a new vertical gpu mount coming, and I’ve also ordered a display port repeater/cable for my HP Reverb G2 as not sure if Nvidia/HP have sorted out the display port issue with the 3090TI/Reverb G2.
Card should be here by Wednesday owing to the bank holiday in the UK today. I will update when it arrives
1 Like
I just hope it’s already been confirmed a refund is due. I once had an RMA’ed cell phone returned to me saying it had got wet (it hadn’t), it turned out to be the seller’s standard answer to all RMAs.
Yeah it has been confirmed with the store. Got it in writing that as soon as they receive the broken gpu which should be Wednesday, the refund will be processed and issued. Currys are good like that in the UK.
2 Likes
With 450 watt you can generate two times more heat then with 225 watt.
Therefore the risk of extreem overheating is higher on a RTX3090ti then on a RTX3060ti.
well… it sounded more like a rant about the Gigabyte RMA 
So,… you ask whether other users have same experiance with failing GPUs and about a possible trend about manufactors ?
Yep… thats not realy an apsect from MSFS, like @Gandalfuski mentioned, because a faulty GPU have nothing to do with MSFS. But on other side yes, may be we need also topics which are less MSFS centric ( but may be in special category ? ).
At least, I already wrote what I think… Its rare in my experiance that a GPU fail , but it comes more and more often on top. And here I agree fully with @PliantWings5288 ( more and more power, not related to Gigabyte , It’s just bad luck!, etc ). Sometime users also overcloack these already factory OC’s cards and wonder that strange things happen.
For what you should go ?
Well, the biggest PC case you can find , so that is allways the best airflow which is possible. All others is luck or bad luck. And you were lucky in misfortune , you are in warrenty. The only thing what happens is: you have to wait 10 days
I would highly recommend that you check your internal airflow and temps inside your PC case so that you don’t end up with another dead GPU. On my PC I’m always displaying CPU and GPU temps with overlays. On previous PC’s I even implemented an auto shutdown procedure if temps reached a threshold. GPU’s today are designed to throttle back once temps get to a certain temp but you never know after repeatedly reaching max temperatures.
I’ve a 011 dynamic evo with 10 fans so I’d like to think that covers airflow, what do you think?
This is how it’s set up with the now half dead Vision RTX
Gpu never exceeded 75c. Memory temps never exceeded 85c. My i9 12900ks never exceeds 80c
As you can see I’m the same and always display cpu package and gpu core temp on my elite lcd display
Witch are in and witch are out?
Efficient airflow is not only dependent on the number of fans.
To much fans can brake the efficienty of the flow direction and create unwanted pressure in the case.
3x inlet on the bottom and 3x outlet on top are more then enough to create efficient flow.
Given those temps your golden!
Side in, bottom in, rear out, top out
I like Currys. A couple years ago I bought 65 inch 4K Philips Ambilight TV for £749 in a Christmas sale. Just over a year later the backlight started flickering. I reported it, they collected it a couple days later, said it wasn’t reparable and asked if I would accept a full refund credited to my Curry’s account to buy something else from them. They had a Philips Ambilight 70 inch 4K for £799 in the sale then so I just gave them £50 and got that delivered the next day, with a brand new warranty.
I’m hoping this one breaks at some point so I can return and get a refund towards an OLED panel…
1 Like
I run a gigabyte rtx 3080 with an 850W psu and have no issues. I use gigabyte exclusively from motherboard to gpu to psu and am very happy with the product.
Nvidia strickly recommends a PSU 750W or greater for 3080 cards and higher.
1 Like
For Nvidia 3080 a 750 watt is enough, but a bit more is better.
I prefer not to have a load on my PSU of more than 80%.
And also look to the future many upgrading there configuration with more Ram, extra Harddrive, eventually a faster GPU. 750 watt is not enough for a 3090ti together with a high power CPU. So you have to buy a new PSU.
1 Like