This is not Blizzard’s fault.
We have a Gigabyte RTX 3080 Ti and at least in our case the issue was caused because our card’s failsafe for temps built in Gigabyte’s bios obviously didn’t work as it should, resulting in the card overheating.
Thankfully, our card didn’t die.
It would have died for sure at some point if we hadn’t been monitoring our gpu temps regularly and kept playing games without knowing our card had been overheating.
It was just a matter of time till that would happen.
Since we bought this Gigabyte RTX 3080 Ti about 3 months ago, we have tested it at full speed under full load (99% usage) in dozens of very demanding games and synthetic benchmarks, and the card would always run at about 1860MHz boost speed and 73 degrees Celsius with the fans spinning at around 70-80%.
For reference, here’s a video of our 3080Ti benchmarked in several demanding titles and as you can see, the card never exceeds 73 degrees at full speed, under full load:
Diablo 4 was the first game to ever exhibit such an abnormal behavior and this happened the second time we run the game. During the first day, we played for hours without an issue.
We have this anomalous behavior recorded in this video:
Diablo 4: We Recorded the Mysterious Anomaly Killing GPUs with a Gigabyte RTX 3080 Ti:
We had never seen something like that, this just didn’t make sense. One of the commenters in the comment section suggested that this might be caused due to an overloaded memory rail and even posted a technical video explaining why this might be the case.
After the first open beta weekend, we went back to playing various other games, like Hogwarts Legacy for example and the issue didn’t occur since. Another commenter suggested that this might be a fan related issue, but at first we (wrongly) didn’t believe this might be a possible cause.
Then we watched the video again, and noticed that we had forgotten to set MSI Afterburner to display the card’s fan speed while running the game. So, we went in the settings and set the fan speed to be visible in the UI.
A couple of days later, possibly out of coincidence, the card started exhibiting the same abnormal behavior in other games as well and not just Diablo 4. But this time we had one more clue that made everything obvious: the gpu fans were running at just 41%, which is way too low. In fact, as we found out later after testing, the Gigabyte RTX 3080Ti has fans that practically cut off completely at 41% speed.
In plain words, for some unknown reason, even if the card was under full load, the fan controller was acting as if the card was still in idle state, so the fans wouldn’t spin as the temps got higher and higher and the gpu was throttling, decreasing its operating frequency all the way down to just 300MHz, trying to avoid frying itself due to overheating.
We installed the card in a different system to test it further, and the issue didn’t occur on its own, so we tried to replicate the issue artificially by setting the gpu fan manually to run at 41% and voila: The abnormal behavior occurred again!
So, as it turned out, in our case the issue is related to a fan controller malfunction resulting in the card overheating.
After some more digging, we have found that the culprit is most probably the Gigabyte bios for the RTX 3000 series AT LEAST PARTLY, and while the issue seems to trigger as a combinational result of several factors (possibly due to software conflict with some other program, like rgb software, etc), it is most probably related to at least one or more Gigabyte bios versions for the RTX 3000 series cards, as it is known to cause fan issues in some cases, here is a reddit post from 3 years ago for reference: https://www.reddit.com/r/gigabytegaming/comments/jv7bcr/gigabyte_rtx_3080_gaming_oc_fan_issues/
Gigabyte sadly haven’t released an updated bios version that resolves the fan issues with their RTX 3000 series cards.
The solution to the problem is either applying a custom fan curve with MSI Afterburner before running a game, thus bypassing the Gigabyte bios fan curve (recommended way), or flashing a compatible BIOS for the same card from another vendor (eg ASUS) on your Gigabyte card (unorthodox way).
If you have a Gigabyte RTX 3000 series card and you’re facing gpu overheating issues randomly, that seems to be the only workaround, at least for now.
However, there have been people reporting that while they had been constantly monitoring their gpu via MSI Afterburner with a custom fan profile applied, FPS capped at 60, and GPU never exceeded 50% load and 60 degrees Celcius temperature for the majority of the time, it still crashed and got bricked:
In cases such as this, hardware getting bricked without having overheating issues is obviously faulty, some component on the card just fails and the card dies.
This means that we most probably have two separate things, sometimes occurring at the same time:
#1. Some cards (from any manufacturer) can have defective components (that’s nothing new or special)
#2. At least one or more Gigabyte bios versions for the RTX 3000 series cards reportedly has a faulty/buggy bios fan curve causing fan issues that in some cases could potentially result in gpu overheating
If you have a card with issue # 1, your card may fail at any given time when stressed. Nothing new or special about that, that’s why we have RMA.
If you have a card with issue # 2 (that is a card from Gigabyte with a faulty/buggy bios fan curve), there is a possibility that even if the card is under full load, the fan controller might act as if the card is still in idle state, ignoring the readings of the temperature sensors, so the fans won’t spin as the temperature increases and the gpu starts to throttle, trying to avoid frying itself due to overheating.
This is what actually happened to our Gigabyte RTX 3080 Ti and as we already mentioned, we have this anomalous behavior recorded in our video.
Thankfully, our card survived this anomaly, hopefully because it didn’t have issue #1 as well?
We can’t say for sure, we can only make an informed guess.
But as it is evident, if you have a card with issue #1 and issue #2 occurring at the same time, chances are that your card won’t survive the anomaly, as the extreme stress in combination with overheating will obviously put a huge strain on the card that will make any defective component fail sooner rather than later.
We sincerely hope that the companies involved will soon be able to identify and resolve the issue once and for all.
Until then, ALWAYS, and I mean ALWAYS, monitor your cpu + gpu temps with MSI Afterburner while playing games. After all, this is something that we highly recommend every PC gamer to do at all times.
Here is our guide on how to download and setup MSI Afterburner:
In short, in order to be able to play this game (and any other game for that matter) without issues and without worrying about possibly damaging your gpu, ALWAYS monitor your gpu temperature, usage, frequency and fan speed at all times with MSI Afterburner and set a custom fan curve before running your games.
Also, remove the Diablo IV Prefs file from the Diablo IV folder in the Documents location.
Normally you shouldn’t have to do all this just to be able to play a game, but as I always say, if there is something on our end that we can do to save us time, money and frustration, we should definitely do it.
Better be safe than sorry. 
Hope this helps!
Anthony,
Palatech Gaming YT