Larian Studios Forums Baldur's Gate III Baldur's Gate III - Technical & Gameplay Problems BG3 shutting down PC entirely

My PC has started to randomly power off whilst playing BG3, and often doing something completely mundane like talking to an NPC, sorting my stash or loading a saved game.

I have to switch the PSU off and leave it for a few seconds and back on again to even get the power button to work again. This is in DX11 mode but same thing happens in Vulkan.

Strange thing is it's only BG3. I've tried Cyberpunk and Diablo 4 both on ultra settings and I can play them without triggering a single crash.

I've used afterburner to verify my CPU does not go above 80 or anywhere near 100% usage when these crashes happen, GPU is chilling in the 60s

I've run repeated stress tests in cinemark and Heaven to try to trigger the issue, will not trigger. I work from home 3 days a week and my PC can be on all day with no crashes whatsoever.
1 keep the PC clean and well ventilated, all my drivers are up to date. GPU driver installed fresh with DDU in safe mode to remove older drivers before. Nothing is overclocked in BIOS.

Specs:
AMD RYZEN 7 5800X
GTX 3060 TI
32GB DDR4
850w Coolermaster 80 Gold rated PSU
AData XPG Gammix 1TB NVME Gen4 (OS installed here)
Crucial MX500 2Tb (Game installed here)

I'm concerned if this keeps up I'll do some damage to my components.

Tomorrow I'II try taking it to bits and reassembling (not sure this is the issue) and swapping out the PSU for a Corsair 850w one. I really think this a BG3 issue that's going untreated though, and not my PC hardware. I'm seeing scattered posts across Reddit, Steam and Linus Tech Tips with people having the same issue.

you've already found the problem... it's most likely your PSU. got a spare you can test with?

Today is now tomorrow and before taking it apart I'm testing the
RAM with Memtest86 (currently on pass 3 no errors found).
If the RAM is fine I'll use crystal Disk to test the SSDs next, then if they are okay I'II completely delete BG3 and reinstall from scratch. I have previously verified the file integrity through steam and it found no issues, even cleared the cache in the APPDATA Larian folder.

Something I'm not seeing here and disassembly / new PSU is last resort will update when I finish testing on whether I can get it working.

Not entirely sure but I think Saving/Loading may be a trigger in there’s critical system error as I can usually replicate it by repeatedly quick loading/going to game main menu and loading.

Not sure why these crashes would only happen during BG3 sessions though if it is the PSU - I’d surely be getting them elsewhere too especially whilst stress testing or playing other intensive games?

I don't agree with the other commenter about it being the PSU; an 850W PSU should handle those specs easily. There are dozens of things it could be,though, and going through each and every one will take tons of time. That said, one could e.g. start with downclocking the CPU, GPU and GPU's VRAM and system RAM, move the game to the other SSD and then seeing if it still continues to do it or not. If the issue disappears, then proceed to turn only one of those things back the way it was at a time and seeing when it begins to crash again.

First I would make sure that this is only a BG3 issue by putting in a couple of full sessions on at least two other games.

If it turns out to be happening with others games then I would physically check inside the case to make sure that fans (including the PSU fan) are working and that heat-sinks aren't clogged up.

If it does turn out to be just BG3 then I would make a careful note of where and when the crashes happen and how long I had been playing for. If you can reproduce a crash either by time and place or by length of game play then send a report to Larian.

100% not their PSU, im having the EXACT same problems, even after deleting the game and all the files in app data every few hours my screen starts freezing and my computer needs a total restart to get back into working condition again.

Like....Jesus Christ this game is already a bug-riddled hot mess already that made me unable to finish act 3 due to how broken it is and now ontop of that its completely ruined my GPU?

A quick google search will tell you that there are lots of people whose systems have also been permanently damaged by this buggy mess and it pisses me off that now im going to need to spend a ridiculous amount of money buying new parts for my system all because Larian decided that they had to beat the starfield release and pushed this bug-filled mess they call a game out way before it was ready.

I have some issues too. The game starts to slow down and then characters start disappearing quickly with their ears and weapons left in the air with nothing else. Something is VERY wrong. It did not destroy my GPU thankfully, but something is scuffed here.

If your system is powering off (as opposed to getting a blue/black screen of death, which usually indicates a driver fault) then the problem isn't with Baldur's Gate 3, it just happens to have pushed your system hard enough to expose a hardware problem.

First thing to check should be temperatures; install HWInfo or some other tool that lets you monitor temperatures and see how hot your system is getting after a while (you may need to use GPU software to get the GPU temperature). You'll want to double check what your hardware's tolerances are, but in general you don't want a modern CPU going over 90ºC, GPUs can usually handle a bit more, but ideally you want them cooling to below 70ºC if the cooling is doing it's job. But again, check specs for your specific parts as some do just run hotter and that's fine.

If you think your system is getting too hot, then as Beechams says check your system isn't clogged with dust and the fans are running. HWInfo and other monitoring tools can tell you what your fan speeds are like, but unless you only have two fans (one for CPU and one System fan) it probably won't be able to give you speeds for them all, so you need to check air is flowing as expected.

If it's not the temperature, then the next to check would be any overclocking you (or your system provider) might have done; this is much more complicated, but overclocking (and undervolting) can lead to instability that can cause sudden restarts when a CPU/GPU fault occurs as a result. You'd need to check what the default speeds are for your CPU, GPU and RAM so you can compare to what they're reporting as, to see if they've been altered. If something has, try putting it to default and see if that solves the problem; this may require you to poke around in BIOS settings.

Only if none of this helps would I assume a faulty PSU; though in my experience a faulty PSU will usually fail randomly or soon after startup, rather than waiting until you put it under load.

Okay so after a day or so of testing various components with benchmarking tools this is what I discovered:

Prime95, Cinebench, Heaven Benchmark, Memtest86 and monitoring with HWInfo64, HWMonitor, Afterburner and even CrystalDisk for checking SSDs - none of which crashed my Pc during testing and none that showed any signs of corrupt memory or temp spikes.

I was on the cusp of changing the PSU, when I noted that the voltages looked fine in HWMonitor and never send to fluctuate throughout testing, so to me looked stable.

I decided in a 3 hour session of Diablo 4, but this went by without a hitch and did not crash the PC.

Stupid I know, but as a last resort before ripping everything out I decided to delete BG3 completely (excluding saves that I backed up), and reinstall it.

As I booted up steam before I uninstalled BG3 I did notice a small patch (after the major one downloaded) but I deleted the install all the same and then started the download over.

Whilst I did this I cast my mind back to the time of the last shutdown/crash and trawled through the event viewer log. I found immediately before the “6008 error the system shut down was unexpected” (created when I booted up the PC after the crash) that there were 3 entries of a particular registry key requesting local activation permission for the COM Server application with CLSID xxx and APPID xxx that permission was not granted.

I took a hell Mary and decided once BG3 (if this was involved any way) reinstalled I’d create a desktop shortcut and always right click and run as administrator.

I’ve now had 2 5 hour sessions with no crashes, leading me to believe:

The delete and fresh reinstall could have fixed it.
The “run as administrator” could have fixed it.
The random Larian update could have fixed it.
It may not be fixed and will reappear later.

Either way I wanted to update you guys. I haven’t replaced anything yet but as a reward (even though it didn’t need it) I cleaned the PC and all fan blades today to keep it happy.

I’ll let you know if it happens again, if it doesn’t I hope what I’ve wrote helps one of you also having the issue.

i assume you have a vpn or firewall in place and it sounds like the issue was your network which would make some sence if you also have a tick in cross saves

anyway glad you got it sorted

I had this in EA. Could barely get past character creator.

I also had no problems with cyberpunk, horizon zero dawn and other demanding games at the time. Disco Elysium on the other hand was equally unplayable.

It was definitely hardware for me, but a new PSU didn’t do the job. Eventually, I’m afraid to say, I upgraded everything, but my system was getting on a bit anyway. Hopefully you don’t have to go that far. Good luck.

Having the same issue. I swear its bg3. This started after the latest update. I have
*SPECS*

Intel i7-13700K

Corsair iCUE H150i RGB ELITE 360 Liquid Cooler

ASRock Z790 Steel Legend WiFi

Corsair Vengeance RGB 64 GB (4 x 16 GB) DDR5-5600
CL36 Memory

Rtx 3080 evga

Lian Li 011 Dynamic EVO ATX Mid Tower Case (Black)

Asus ROG THOR 1200P 1200 W 80+ Platinum Power Supply

Every other game works fine, no crashes etc. I also ran a stress test called OCCT for 30mins. And unistalled and reinstalled.

This game turns off my computer when i engage in dialoge or go to fast in some actions with my controller.

I have the same issue.

I have a very similar system with the poster above, jedi97yey. If someone finds a solution please post it, also OP please let us know if your issue has not reemerged.

Personally, I tried everything systemwise before finding this post (and the other in steamcommunity) including replacing my PSU (with my old one) as I was convinced it was my PC's issue. I have not tried SinisterCP's solutions (reinstall/run as admin) as I am afraid that if I start the game I may damage my PC.

Had the same issue. Ryen 7800x3d, rtx 4070, 850w psu. When I loaded this a save or entered a video, randomly, oops PC shuts off. Cpu temps high but ok on overlay so if the spike happened it was fast. No GPU temp issues.

Solution was to go into overclock settings in bios and put a max on the temp for the chip. I put it at 65 no issues now. I'll try 75 or 85 maybe after a few weeks. System seems to be limiting clock speed on cpu to keep it in check. Performance impact negligible.

Others I have read achieve the same by limiting the wattage on their CPUs in bios or by enabling eco mode on certain motherboard/cpus.

I tried eco mode in ryzen master. It did not work. Only bios did for me.

This leads me to believe something going on with game causing cpu drawing too much power and/or hitting high temp. Probably the former as there did not seem to be any metallic smell and all other temps were great.

I just wanted to let you all know the issue still has not resolved, it’s completely random for me and I’ve not played much to test it as I’ve been in process of moving this week - the last time I booted up the game it shutdown my PC simply loading a quick save.

Frustrating, I’m changing out a couple of components this weekend - PSU for an ATX 3.0 model same wattage (850w) and a new WiFi card not that either of them are a problem anyway as other games have been fine, even cyberpunk which I would go as far to say if it was a cpu/GPU issue that should show similar issues.

I honestly think with this many of us having the issue it has to be something to do with how the game is optimised that Larian need to fix on their end. It’s heartbreaking because I’d really like to play this game and now I can’t.

I am having the same issue and my power supply has been replaced and it happens exclusively when playing this game, but not randomly (It crashes when I walk over certain spots or long rest, but thats a whole other tale). I totally agree though, the game is unplayable like this and that is really annoying

It's definitely a BG3 issue and I am confident that Larian will eventually fix it. It seems that we are all experiencing the same crash, and it's a critical one - very unlikely that all of our PSUs are suddently faulty. If you haven't done already, please submit a bug report.

Personally, I followed Kbottt's advice above and my issue has improved/resolved.

MY TEMPORARY FIX:

At the beginning I was getting this crash after 30 minutes to 1.5 hours of gameplay. It was happening during dialogue, combat or even character creation. It happened thrice before I started tinkering with solutions, including updating drivers, lowering gpu voltage (alt+z), reinstalling game and changing my PSU entirely. In total I think it happened around 12 times before I stopped trying. In order to protect my PC.

A few days ago I tried Kbott's solution - went to bios and put a power limit to my CPU (also I disabled all stock "overclock" configurations since I have a gaming z170 mobo). After doing that I was able to have an 8 hour session, before I had another crash. Originally, I was rarely passing the one hour mark so I was/am hopeful that this can be a temporary fix. I then went back to the bios and undervolted the cpu even more (basically I chose the highest power limit possible) and since then I had 3, long, uninterrupted sessions. I will edit this post if I encounter a crash again, but for now the game is playable.

More detail:

Similar to Kbott above, my cpu temps were and are fine on the overlay - so if a spike is happening it is very fast. I have a 12700k and the Be Quiet Dark Rock Pro 4 which is considered one of most effective cpu coolers, so if the issue is actually temp-related I don't think that upgrading/changing your hardware will help. Furthermore, even with the most extreme undervolting option I saw no/negligible performance loss, I am steadily at 60 (frame limit) frames in 4k. However, I also loaded my ultra-modded skyrim installation and the fps reduction was ridiculous, so if this fix works for you too, you should expect that you will have to go back to the bios and change the values for other demanding games or work. Finally, I also removed the power limit of my GPU a few days in and the crash did not reappear, so I suppose it is unrelated to the gpu.

Still happening for me and it happens whenever I do certain actions in game. Walking across the bridge between Emerald Grove and The Blighted Village, Walking around the area South of the Blighted Village, talking to the goblins outside the Blighted Village, talking to Karlach the first time in the wilderness, walking into the goblin camp west of the Blighted Village, and after taking a long rest (I usually see the auto save right before the crash).

I can reliably recreate this crash every time by doing any of these actions and even making a new character it will still happen.

I have tried adjusting every setting that is suggested, tried both DirectX and Vulkan, bypassing the launcher, and grasped at every straw I could find anywhere and am at the end of my rope. I am not sure if this is a problem because I am running an AMD chipset, but temperature monitoring is fine and as mentioned above the power supply has been replaced all on a 5 month old computer so I am certain it is not hardware failure at this point. I have sent a bug report and some requested files in to the automated reply, but honestly I dont have any faith whatsoever that they'll even be seen by a human much less addressed.

At this point I am disgusted and disappointed customer because I can't even ask for a refund at this point all I can do is wish the game worked

I have the same issue with BG3 and with Starfield on my new installed GeForce rtx 3060 but when I replace it with my old GeForce gtx 1650, the game runs (bad of course because of crap video card) without PC shutting down. I believe this is because my old cpu Intel 3770 doesn't generate enough juice to feed my rtx 3060. My 700w psu is more than enough for my system

Same solution for me. 75 also works but limiting to 85:still caused shutdown.

This was with cpu fan on max and smart.

Curious if your psu has silent fan mode? Button on this was pushed in, meaning maybe my psu overheated? Butnwhybjot starfield....

Were you able to find a solution to your problem?

I'm encountering the same thing as you are. AMD 5600x, 3070, Corsair 750 Plat. All other games run with no problems, except BG3

I can alleviate the problem by running the game at low resolution, window mode, and low settings, but if choose default graphics settings, at 1440p, the game will suddenly turns off my PC

Hi there, I was going to open a thread to talk about it, but decided to reply here. The same thing is happening to me. My computer shuts down out of nowhere while playing BG3, it doesn't overheat, I tested harder and stronger games on max and nothing of the sort happened. The PSU is fine, it doesn't get over 70° neither in my CPU nor my GPU, yet out of nowhere while playing the game, the pc just shuts down. Toning down the graphics only delayed the inevitable, and sometimes, things just happen at random: One session I played hours with no problem, others it was nearly right after starting the game. Sometimes all it took was asking for a character to sit down, others it was during an intense cutscene.

I tried nearly everything, from undervolt modes for the GPU to energy saving modes for the PC, nothing works.

Anyone found anything about this problem?

I'm also experiencing the same issue. Hoping someone figures something out. Most of the replies out there are "fix your hardware". I'm just so skeptical that it's a PSU issue across all of these different machines and set ups, completely and solely specific to BG3. It's definitely not temps. I dunno, but I'm super bummed since I don't want to risk my hardware to play this very enjoyable game. Larian, halp!

Sorry to hear that. And for everyone else having this issue.

I’m no expert, but a hardware issue of some kind does appear to be at least part of the problem. As I understand it, an emergency shutdown is a hardware failsafe to try to prevent damage to the system (even though repeated resets isn’t particularly good for it either). Programs can lock up, crash, cause fatal errors and all that, but I don’t think they can just pull the plug without something else being wrong.

When I had this problem very badly, it was not just BG3. Disco Elysium for example was completely unplayable as well for me. Other games and programs were generally OK for the most part, but the problem was still there. On cyberpunk 2077 for instance I still got the occasional hard reset, maybe once every 5 or 6 hours. So it’s feasible that others could only see the problem on BG3.

My hypothesis is that there’s something about BG3 (and DE) that taxes some part of the system more than most other games and programs, exposing a weakness. In my case I don’t think it was the PSU (tried replacing it) or temperature (didn’t look too excessive), but I’m not entirely sure.

The fact it’s happening across different setups doesn’t rule out a hardware issue. Components from different manufacturers all do the same thing and can all fail in the same ways. This doesn’t entirely let Larian off the hook though. If there’s something in the program causing issues in systems that seem fine running most everything else, they should definitely investigate and improve it.


It doesn’t help that a hard shutdown doesn’t generate crash reports.

Yeah, at first I thought it was my PSU, but then I saw people with way better rigs than mine having the same problem. I uninstalled the game and I will wait for results, I'm also afraid of BG3 damaging something on my pc. I have faith that Larian will discover what the issue is and fix it.

A system shutdown is almost always an indication of the hardware issue. This could be a current over drawn, a heat issue in the PSU / CPU or GPU. (Could also be motherboard) I have never in over 30 years as a tech seen a program be the cause of this kind of an issue. I can be the trigger, but not the cause.

Hope can you, but Larian cannot in any way fix your shutdown problems. Shutdowns are 99,9% caused by your hardware.

What Larian can do is optimize their game to use less resources and then your hardware problem can go hiding again. But they never can fix it. Computers need to be anle to work under 100% stress and not shutdown.

My estimation of your problem is not psu, but motherboard vrm.

When discussing these problems people should give exact specs, names and models of components. It is farily easy to spot subpar motherboards for example that can not run cpus…

P.s. I am willing to bet decent sums of money that your computer cannot run hours of lets say prime95 and msi kombustor at same time. Those software put your cpu and gpu to 100%.

I would love to have an answer to this, even a common thread on hardware everyone with the issue is using. It's just hard to find the time to take some action without a more serious handle on the problem.

I wish I had the time and money to start swapping out parts to see if I could isolate it, but I'm an amateur who games in his free time. I'm really not trying to assert anything except I'm having the same problem that it looks like other people are having. The path of least resistance for me is to just keep eyeballing forums like this and to move on to a game that doesn't crash my system, i.e. every single other one I've tried so far.



Thanks for this response, very diplomatic and you articulated what I was thinking very well. Here are my system specs for the hell of it.

AMD Ryzen 5 5600 3.5 GHz 6-Core Processor
ID-COOLING SE-214-XT 68.2 CFM CPU Cooler
Gigabyte B550M DS3H AC Micro ATX AM4 Motherboard
Kingston FURY Beast RGB 16 GB (2 x 8 GB) DDR4-3200 CL16 Memory
MSI MECH 2X OC Radeon RX 6700 XT 12 GB Video Card
Patriot P310 480 GB M.2-2280 PCIe 3.0 X4 NVME Solid State Drive
Intel 670p Series M.2 2280 2TB PCI-Express 3.0 x4 QLC Internal Solid State Drive (SSD)
Apevia Prestige 600 W 80+ Gold Certified ATX Power Supply
Microsoft Windows 11 Home

ATI AMD Radeon RX 6750 XT
AMD Ryzen 5 5600G
Corsair CX750M for power.
All are new.

24,0GB Dual-Channel DDR4 @ 1064MHz
447GB KINGSTON SA400S37480G

One potential issue is your running mixed RAM. Looks like your running 2x 4GB sticks and 2x 8 GB sticks. This is usually problematic and can create stability issues. Try dropping to 16Gb (dual 8Gb) and see what happens.


Look at your list two components leap out at me. The Motherboard is the first, if the system has decent air flow it will likely be okay but is a potential weak link. The second is the PSU, Apevia is not a well respected brand and this could be the weak link in the system.

You did not list the case and fan setup, this can be an issue if the system is not getting great air flow. There is an easy and quick test to see if the air flow is the issue, take off the side of the case and run it without the side panel. If the system runs as it should without the side panel then the issue is case air flow causing a heat issue, likely on the VRM.

Before you do that however make sure ANY overclocking is turned off, you want to test this at stock.

This. Extremely unoptimal memory setup. Plus mobo is not even listed.
Also PSU is very low tier. I bought one cx-m unit for my moms pc and it bbroke down immediately. For myself I would never consider.
I agree. This motherboard has one of the worst b550 and b450 VRMs hardware unboxed tested. If i recall correctly hwinfo64 shows VRM- temps on gigabyte motherboards. That should be monitored.


And like you said. Apevia psus are considered to be in avoid list as they have been horrendous garbage. This exact model info is bit limited at this time. But I would never ever buy such power for anything.

It's great to have some tech-savvy ppl responding to this thread. Thanks for sharing your knowledge. I was having a similar problem with my PC and after doing a fair bit of digging (I am not, by any stretch of the imagination, proficient with the tech side of computers; I merely dabble), I came to the conclusion that I needed to do three things, in particular. First, I needed to reset my computer to its factory settings and reinstall my OS (Win 11 Home). Second, I needed to update my BIOS. Third, I needed to update my chip set drivers. After doing these three things, my BSOD crashes mid-game have completely stopped. However, BG3 *does* still crash all on its own fairly often (more often than a game should), but it no longer crashes my entire system.

I might draw the ire of AMD enthusiasts but I have a sneaking suspicion that these crashes are perpetuated by AMD parts. I've always had Intel/Nvidia systems in the past and decided to give AMD a try with my current rig and it has given me nothing but regrets.

I'll list my build below, in case someone much smarter than me can see something wonky about it.

AMD Ryzen 7 5800X Vermeer 3.8GHz 8-core AM4 Processor
MasterLiquid ML240L V2 240mm water cooling
MSI x570S MAG Tomahawk Max WiFi AMD AM4 ATX motherboard
Corsair Vengeance Pro 32GB (2x16) DDR4-3200 PC4-25800 CL 16 Dual Channel memory
MSI AMD Radeon RX 6800 XT Gaming X Triple fan 16GB GDDR6 PCIe 4.0 graphics card
Samsung 980 SSD 1TB M.2 NVMe Interface PCIe 3.0 x4 solid state drive
PowerSpec 850W 80 Plus Gold ATX modular power supply
Lian Li Lancool II Mesh Type C ATX mid-tower case
Microsoft Windows 11 Home

You have for example good cpu acompanied with top tier motherboard. Ssd etc work well.

But then you have acompanied that with low tier psu. Brand that have made absolute garbage psus. Maybe your is not, but it is low tier at best.

Another thing to note. Amd ryzen benefits cpu speed wise from good memory. It’s internal parts run speed relation to memory speed. I would consider your memory speed ”minimum” reasonable for ryzen to operate properly.

Not all ryzens can run 4000MHz memory (mine does not), but 3600MHz memory speed (and thus ryzen internal bus speed) should be everyones aim.

So you have mixed top tier parts (mobo+cpu) with (lowish) mediocre (ram) and with low tier parts (psu).

In case of shutdown problems I would walk to store and get PSU I can trust and go from there.

For those impacted using an AMD CPU: worth checking that your graphic card is using a PCIe slot connected directly to the CPU and not one connected to the motherboard chipset. Generally, it means checking that your graphic card is in the topmost PCIe slot of your motherboard. Also, it may be worth testing with every other PCIe devices removed (if any) (NVMe drives using the topmost slot should also be fine).

I have encountered instability issues with how some AMD motherboard chipsets handle high load. Strongly doubt that this is what's happening for most people here though, but who knows.

Edit: also, I think someone else already mentioned it in this thread, but maybe trying to update the motherboard BIOS ? I'm bit reluctant to recommend it, given that it's a bit of a crapshoot whether it will actually help or make things worse.

For owner of AMD GPU cards, I have heard that updating the GPU BIOS can help. Never tried it myself though, and it doesn't seem to be officially supported by the manufacturers. Feels like a bit of high-risk procedure.

FWIW, since the most recent patch, BG3 is also causing PS5 shutdowns when the game client freezes and crashes. I'm not sure this is just an end-user PC hardware issue.

I have three systems in my home, all running BG3 for family members and none of them have an issue so this is not an "AMD" issue. This could be background software related issues.



NEVER update a GPU BIOS unless directed to do it by the GPU tech support team.

My computer shut down twice today while playing BG3. Exactly as OP said, I have to switch the PSU off and leave it for a few seconds and back on again to even get the power button to work again.

I've playing almost everyday since launch, and it has never happened until today. Maybe the most recent patch has something to do with that in my case as Erica pointed out.

Hardware:
Ryzen 5600x
RTX 3060Ti
600W PSU

This kind of issues are always caused by hardware, not software.

It is your PC, the game just happens to be enough of a stress test to show it.

I've got a RTX 4090 and i9 and getting these shutdowns too, actually so disgusted by this game

I have witnessed a Sapphire RX 6800 XT Nitro+ that would reliably crash the system the moment its junction temperature reached 110°C (not 109, not 111, always precisely 110). The initial symptoms looked a lot like what's described in this thread: most games were totally fine, but some games (the ones pushing this particular part of the GPU a little further) would cause a system crash after some play time. The issue would also appear if the card was moved into a system that was perfectly before.

I concede that this is not a hard proof of an issue in the GPU BIOS (we could not attempt an update to test with a newer version). But given how repeatable the issue was, and the fact that the card was part of one of the first production batches, I remain convinced that this was a bug in the thermal throttling mechanism.

For the chipset issue, I have also been able to fully and reliably reproduce the issue on completely distinct B550-based configurations on Windows, Linux, and FreeBSD (also tested with various BIOS versions). That was about 7 months ago.

As much as I want to support the underdog, gotta say I have not been impressed with AMD's QA those last few years.

Leaving a PC shut off like that to get it to come back on is either a heat issue or a PSU.


This is a fairly high junction temp and sounds like a cooling issue. Try removing the side panel and see if the card stays cooler (want to see around 80C or so for junction, below 90C is okay. If the card is till getting to 100C plus on the junction contact manufacturer support as it sounds like the card is having a cooling issue.

The shutdown could still be from a PSU, as a CPU or GPU hits these higher heat numbers they can have increase power draw and it could be the PSU is not keeping up. However I would first look at that GPU temp.

From what I read at the time, 110°C for the junction hot spot was fairly common for factory-OC cards and not considered an issue by AMD. My own 6800 XT that I had at the time occasionally reached the infamous 110°C and just throttled down without issue. It was from a newer production batch though (thanks the chip shortage crisis for that !). Also noteworthy; the other sensors of the card would stay within the 70°C - 80°C range.


I don't think so. During our tests, we moved the card into my own system where it would exhibit the same behavior.

In the end, we "fixed" the issue by installing 2 extra fans blowing directly on the card.

If your having to add fans to blow directly on the card the issue is air flow.

That really tells that the GPU has HW issues and/or AMD has drivers issues. Cooling the card makes it not crash. Proper behavior would have been automatic throttle like you mentioned.

So it would be AMD and/or Sapphire issues you originally claimed I guess.

Hi this issue has been happening to me for a couple of months now. In trying to fix this issue I've changed a lot of hardware and the issue has seemed to persist.

Build is currently:
-AMD Radeon RX 6750 XT
-Ryzen 7 5800 X3D
-32 GB Corsair (2 16GB sticks)
-850W Corsair gold PSU
-ASUS TUF x570 plus wifi motherboard
- 850W Corair gold power supply

Initially I was using a 750W PSU and changed it out for another 750W PSU. My experience was that the shutdowns stopped for roughly 3 weeks before becoming a daily issue again and onyl with Baldurs Gate 3.
I eventually moved up to an 850W PSU which once more after several weeks began to experience the same shutdowns. Usually after 1-4 hours of BG3. This occurs whether my settings are at low, medium, or ultra.
In desperation over the past week I have tried a new motherboard (Same model but brand new board), New RAM, Upgrading to Win 11 with a clean install, and uninstallign and then reinstalling my GPU drivers.

I can't fathom what could be causing the issue at this point. Temperatures do not exceed 72-75 degrees on the CPU nor 77-80 on the GPU end. No power kernal errors in the event viewer following these shutdowns.
Please can someone help with this issue?

@proyal upping your power supply only helps if there is enough actual power so if your mom is turning on the microwave to cook and your sister is running the washing machine there may not be enough to also run your gaming at the same time

Thank you for the reply. I don't think I'm short on wattage however. Wattage calculators put my system between 600-699 watts. At 850W I should have more than enough power, and when I've asked others for help on the matter I've been told my 850 PSU should definitely be more than enough.

To Those who still have this problem i suggest :

- Run BG3 on Administrator mode

- Use ThrottleStop app , set overall as low or battery , disable turbo and speed step on the 2 checkboxes. WARNING : setting ThrottleStop with Performance mode WILL cause the computer to turn off in BG3

Hi
This is not hardware issue given the fact that I'm using an Intel i5 12600 + 5700XT AMD and I get these shutdown.
There are not reboot but clean shutdown however as mentioned before unexpected (windows event logs)
I've for my part be able to fix it.
* I've uninstall the last windows update it was (2023-10 Cumulative Update for Windows 11 for x64-based Systems (KB5031358))
* flash my bios for mainboard (z690 gigabyte) to last version
* update AMD driver to last version 23.10.1
* Run windows check for system integrity see here https://www.dell.com/support/kbdoc/...system-file-checker-in-microsoft-windows (this found error and repair)

Since this no restart anymore.
Hope it can help

Software cannot shut down a computer. It's impossible. What is happening is you are overheating. The same would happen with any other modern game that you run. You need to resolve why that's happening (it'll be a hardware issue). Check that all your fans are working, check that you don't have exhaust fans pulling air in instead of pushing it out, make sure your case has enough airflow for the hardware you are using, check that your heatsinks are properly seated, etc.

Use Throttlestop to stop your cpu / gpu running wild

look at the left side of the app and disable speedstep and disable turbo

this will help

https://www.techpowerup.com/download/techpowerup-throttlestop/

Sorry but this is just not true. Software shuts down a PC all the time, how do you think Windows can shut down a computer? Over the years I have seen Windows cause system shutdowns through file corruption as well. It is not common but it can and does sometimes happen.

Windows is a special case as it's the OS. A video game cannot and will not ever be responsible for your PC shutting down. That is always a hardware issue (overheating, insufficient power, improperly seated components, dying hardware, etc.).

These posts are common on the forums of every game. They are always hardware-related.

also one thing ppl who trouble shoot computers sometime missed ..

- Cables (HDMI , Power extension . VGA) , Power Plug extension , unstable power (not using stabilizer/UPS) sometimes contribute to these random turn off..

- i would replace HDMI cables with higher quality cables for starter , not necessarily the most expensive but make sure it support the 4K

- also check if you used too much device on a single power chord extension , a cheap power extension unit can be bad for electric stability

I managed to resolve this issue by replacing the PSU ultimately. I decided to go for an ATX 3.0 model: Corsair RM850e.

I know the ATX 3.0 PSUs have better management of power fluctuations for the GPU so I don’t know if that’s why it’s fixed the issue completely but I haven’t had a single crash/power down since I did the upgrade (and it’s been a couple months since I originally posted!)

I also note I’m still on driver 537.34 currently (not the latest but I left it as long as possible to be sure it wasn’t something that was patched out of a faulty nvidia driver issue. This was the drivers I was on when the problem first occurred, so it was definitely the PSU at fault - not GPU or drivers.

I can now play the game on max settings with zero issues even if I have twitch and Spotify open on a separate monitor.

Hope you guys having the same issue manage to resolve it also.

Hi all. For those with AMD cards I might have found a solution that has been working for me. I have a Radeon 6750XT and even after swapping in 3 different PSUs the issue occurred. GPU temps never exceeded 77-80 degrees HOWEVER based on some research on other forums BG3 seems to create short bursts of heat at the junction temps. Enough to trigger shut downs. This happens regardless of playing on ultra or even low with all settings down.

My solution turned out to be MSI afterburner. My default core clock speed is 2680 however turning down the core clock speed to 2580 has completely solved my issue and I haven’t had a single shutdown even with all graphics settings maxed out.

You will lose frames turning down the clock speed but it does seem to solve the issue. I would suggest anyone having this issue still to give this a try as it doesn’t require swapping out any hardware.

I want to reiterate that although this is a hardware issue, BG3 is the only game that forces this issue. Even cyberpunk maxed out and in 4K does not cause this issue for my system. Whether this means it’s an optimization issue or something else entirely is well beyond my knowledge.

Yep, I read Proyal's post and in fact, lowered my core clock speed with MSI Afterburner and I've had zero shutdowns since! I was chasing this bugaboo with "Spiderman 1" and had been simply turning off Turbo Boost on my Asus Strix Z790 motherboard. However now, with the core clock lowered, I no longer have to do this and can run everything at full speed with no reboots!!!

I had the same issue. The game was running fine on my old PC, but when I bought a new config, the game started to crash when I tried to load a saved game. If it loaded succesfully first, then it crashed on another load, while other (more demanding games) were running just fine.

I've tried everything, installed Windows 11, then installed WIndows 10 again, fiddled with the GPU overclocking, updating drivers, turning off Steam cloud saves, and everything else I read in forums, yet nothing seemed to help.

But when I set the game's exe (not the launcher) to start in administrator mode, the problem was solved. Or at least it is for now.

My config is:
- Operating System: Windows 10
- Processor: Intel Core i7 12700KF
- Memory: 32 GB RAM (2x16GB Kingston Fury 3200 MHz)
- Motherboard: ASRock H610M-HVS/M.2 R2.0
- Graphics Card: AMD Radeon RX 7800 XT 16GB (Sapphire)
- PSU: Keep Out FX800B 800W