RTX 2080 BSOD in Ark

[Xenithar]

Still no crash for me in a month since i enabled -norhithread, but this is on two 1070's in SLI.vI am using nVidia drivers 416.34, and that is what I installed when I built the PC.

Here's a good question. How many of you did that nonsensical core-parking hack in the registry or elsewhere? Ark can only use two threads so even if you own an 18-core, 36-thread Xeon, core-parking hacks are useless and detrimental. Fun fact. Disable sleep state C6 in your BIOS. C6 is literally the core parking functionality. If you really believe it helps, don't hack stuff up, just disable C6. I bring this up because I have personally seen consumers (you're not a pro because you followed a guide to hack your registry) do these hacks and bring their now broken PCs to me due to BSODs and other things freezing. Seriously, just disable C6 in the BIOS. That is LITERALLY your core-parking functionality.

If you haven't done those hacks and still having issues, are you using that crappy nVidia "GeForce Experience" junk? Try removing it and seeing if it helps.

*EDIT*

Wait, one other thought. Are you SURE you have enough power to run those massive 2080's? If you're doing a 450W PSU I'd bet you're starving the card.

[CyberAngel67]

2 hours ago, Xenithar said:

Still no crash for me in a month since i enabled -norhithread, but this is on two 1070's in SLI.vI am using nVidia drivers 416.34, and that is what I installed when I built the PC.

Here's a good question. How many of you did that nonsensical core-parking hack in the registry or elsewhere? Ark can only use two threads so even if you own an 18-core, 36-thread Xeon, core-parking hacks are useless and detrimental. Fun fact. Disable sleep state C6 in your BIOS. C6 is literally the core parking functionality. If you really believe it helps, don't hack stuff up, just disable C6. I bring this up because I have personally seen consumers (you're not a pro because you followed a guide to hack your registry) do these hacks and bring their now broken PCs to me due to BSODs and other things freezing. Seriously, just disable C6 in the BIOS. That is LITERALLY your core-parking functionality.

If you haven't done those hacks and still having issues, are you using that crappy nVidia "GeForce Experience" junk? Try removing it and seeing if it helps.

*EDIT*

Wait, one other thought. Are you SURE you have enough power to run those massive 2080's? If you're doing a 450W PSU I'd bet you're starving the card.

GeForce Experience has nothing to do with the BSOD crashes or the device being lost.

[CyberAngel67]

It should also be noted that the logs state a Device being lost the moment the BSOD occurs, the device being lost is a year old issue with this game.

[RageQuitter]

4 hours ago, USPfan said:

Definitely not just 2080's. My 1060 in an alienware R13 is BSOD-ing and crashing left right and center. 

Usually while I'm sprinting. The came just freezes. Cant shut down or minimize the window, so I'm forced to restart. Only it wont do that either. I have to power down and reboot, at which point I'll sometimes get the BSOD. I've already reinstalled my geforce drivers. Only happens when I play ark. 

The game just isnt optimized. 

Uninstall the drivers and install 399.24, works fine for 1080 or lower

[thebobmaster]

I think this will never be fixed.

[ADCKnight]

Experienced this for the first time today, i am on a MSI 2080ti gaming trio, been playing aberration and extinction without any issues, popped over to the island for the first time in a year, and BAM, direct3d crashes, device lost, every 2 mins, setting everything to lowest settings and turning off graphics card overclocking made me able to run to the obelisk and upload to safety back to aberration and reverted to max setting and OC, and no crashes, not going back to island again ever i guess.

[XiangZhu]

So I did some extensive play testing this weekend on Ragnarok with a new 2080 (finally got it). Old card had 0 issues with new driver packages (Was a GTX 970).

First night, I didn't do a whole lot. Did my usual upkeep of farm animals, a couple of trips out and about with nothing major to report - thought it was going to be a non-issue.

The rest of the weekend was pretty much hell. Died so many, so many times to BSOD.

Exact same error every time; sometimes would pop up with the error about d3d device lost and then go to BSOD bad_pool_caller, etc. Other times it would jump straight to BSOD.

Inside buildings and areas that are initially loaded into memory are absolutely fine. Not one single problem while in a confined area. The errors occurred when loading in new textures, terrain, and objects at higher than sprinting speeds. This could be on a motorboat or a flyer for me (rafts appeared to be fine?). When flying or motorboating, everything was fine until a large base, or a large new set of textures loaded. I turned settings down on the card and it happened less frequently, but still happened. When re-approaching my base from well outside its render range, it was a guaranteed drop to BSOD if I moved toward it at high speed. If I slowed down and let it render in gradually it would not error. When it had issues, the game would usually studder prior and then ultimately freeze and error.

This error will probably not occur to a dev without actually doing some playtesting on a populated server using a fast flyer or motorboat. Seems to occur less frequently after first loading in, more frequently after being in game a while. I don't believe it's the game itself causing the problem (unless it's an inherent flaw with the game engine itself), more so the driver; but dev support in getting it fixed (aka bitching to nvidia) would be really helpful.

I'll probably do more testing with -norhithread and reduced graphics settings this week - see if I can get a temporary workaround to the problem for myself or identify exactly what to turn off to make it all better in the short run.

[Swordguy]

On 11/11/2018 at 1:01 AM, thebobmaster said:

I think this will never be fixed.

I'm (sadly) going to agree with this.

The reasoning is simple. On the one hand:

1. Nvidia featured ARK as one of the games utilizing RTX features (the machine learning anti-aliasing iirc) when they announced RTX.
2. Studio Wildcard have promoted the RTX brand in the Extra Life stream and probably elsewhere, and even gave a 2080ti away, as per whatever Marketing agreement they have with Nvidia

These two points alone should make the RTX crashing problem a top priority to have been fixed before it became a problem.

 

On the other hand:

1. Nvidia's net profit on older 1000 series cards is likely the same or higher than with RTX, despite what many may think they know about the price increase of RTX. Nobody likes a price increase, and those who don't understand the cost to bring products to market are going to assume Nvidia are just milking consumers. Let's be honest, Nvidia makes their money selling Quadro and Tesla chips. Geforce is the fun department where they get to do cool stuff with the hardware, but its far less profitable.
2. Wildcard already have your money. Anyone who could afford an RTX 2080/ti likely bought the ARK season pass before extinction launched. That just leaves the odd straggler who buys an RTX now to play ARK, without doing their research to see if ARK actually runs on RTX.

 

I hate bashing developers, because the decision to tackle issues often lies with their team leaders, or even higher ups who determine budgets, and I can't help but think in this case that neither Nvidia or Wildcard give two poops about this problem, it just doesn't affect enough of us to matter. RTX is a luxury to own, and until we see 2060's flooding the market, I doubt this problem will see any dev attention.

[Xenithar]

Actually, GeForce Experience attempts to configure settings for you in games and does other dopy things, like updating drivers automatically. You know, drivers which are working FINE as-is. It causes way more problems than it solves. I'll control my drivers and settings, thank you.

Also, for those stating you need drivers below 400.xx, stop spreading false rumors. As I have stated for ages, I am on 416.34 and have been for a while now. Two 1070's in SLI. Zero crashes since disabling the RHI thread. I have played all day a time or two on the weekends. No issues. I even have TrueSky enabled. Blaming the drivers is just plain wrong.

Now for a lesson in BSODs. Did you know that a BSOD is triggered by a hardware error? That's right, kids! No game can tell Windows to blue-screen! This takes a low-level driver which has detected an error, such as low voltage, high voltage, or some other physical error somewhere on your system. This is why I recommended checking the power supply. A single 450W PSU may not be enough to run that shiny new 2080 in 4k with max details. My rig required a 600-650W PSU, so I bought a 750W PSU. Low voltage can cause a BSOD just as fast as an overheated CPU or GPU.

Again, drivers above 400 are fine. If not, please provide solid evidence and don't just throw around a theory you have. Plenty of us are running above 400 with 1070's, 1080's and even the 2000 series without a problem. If the drivers above 400 cause an issue on your system, then you need to figure that out first.

[turbodonkey]

Software  can cause BSOD.

[XiangZhu]

1 hour ago, Xenithar said:

Again, drivers above 400 are fine. If not, please provide solid evidence and don't just throw around a theory you have. Plenty of us are running above 400 with 1070's, 1080's and even the 2000 series without a problem. If the drivers above 400 cause an issue on your system, then you need to figure that out first.

Obviously, if you have to throw in the -norhithread, then you have an issue. Please don't say that you don't. 

Either the devs rewrote the rhi classes of the unreal engine, which I highly doubt, or nvidia made a boo boo and they need to fix it. If they didn't make a boo boo they need to do a work around to fix the 'natural' issue that's occurring since the new drivers came out for the RX cards.

I'm glad the -norhithread argument is working for you, but I'd ask that you not talk down to the rest of us in this thread.

[BluesDS]

I'd also like to add its been shown this doesnt seem to happen with two cards in SLI. I've been lurking this thread for weeks and this comment finally pushed me to speak

[iStephano]

6 hours ago, Xenithar said:

Actually, GeForce Experience attempts to configure settings for you in games and does other dopy things, like updating drivers automatically. You know, drivers which are working FINE as-is. It causes way more problems than it solves. I'll control my drivers and settings, thank you.

Also, for those stating you need drivers below 400.xx, stop spreading false rumors. As I have stated for ages, I am on 416.34 and have been for a while now. Two 1070's in SLI. Zero crashes since disabling the RHI thread. I have played all day a time or two on the weekends. No issues. I even have TrueSky enabled. Blaming the drivers is just plain wrong.

Now for a lesson in BSODs. Did you know that a BSOD is triggered by a hardware error? That's right, kids! No game can tell Windows to blue-screen! This takes a low-level driver which has detected an error, such as low voltage, high voltage, or some other physical error somewhere on your system. This is why I recommended checking the power supply. A single 450W PSU may not be enough to run that shiny new 2080 in 4k with max details. My rig required a 600-650W PSU, so I bought a 750W PSU. Low voltage can cause a BSOD just as fast as an overheated CPU or GPU.

Again, drivers above 400 are fine. If not, please provide solid evidence and don't just throw around a theory you have. Plenty of us are running above 400 with 1070's, 1080's and even the 2000 series without a problem. If the drivers above 400 cause an issue on your system, then you need to figure that out first.

I can agree on that GeForce Experience can create issues. Have seen that many times.

But we need to look at the logic here, 100s of people have the same problem and even get the same kernel crash log. I mean if 2 people got that sure it could be a hardware issue with their PC`s they are using. But we are talking about a lot of people here with different setup that gets the same exact problem. We have all tested the game the same way, like jumping on a flyer on the island map. 5 mins later BSOD.

This problem is in Ark itself as it could be that the game is outdated and can not run 100% on the new card. New architecture and of course Nvidia may have failed at some point. But as i have seen and read this is a software issue.

And dont give me that PSU nonsense i have a 1000w PSU and its giving more than enough power to all components in my PC.

Every other game i have tested on RTX 2080 to this date runs fine for days, for weeks. No windows issues, no game crash, no BSOD, nothing it runs just fine. but as soon i load up the island map with the RTX 2080, boom BSOD 5 mins later. And its always when the game itself load new textures or to be more specific jump on a flyer so the map needs to load stuff in faster.

The new DLC seems to run fine so maybe they fixed something here, but they need to apply that to the older maps as-well.

This is a software issue.

You run on older card with a older architecture thats match with Arks game code better. So whatever you says is invalid sorry to say it. With older cards the driver also works better as you can rollback to a more stable one if something is bugging here. I dont know maybe you play Ark 20mins a day? Some people can spend 3-4h a day and thats when the issue on older cards kicks in.

And another thing, if a person buys a game you can not expect that person to go into BIOS setting to "fix things". It does not works this way. if that was the case most people would stick to console gaming and throw pc gaming out the window.

I fully understand the problem with new hardware and "older games" can crash, its needs some fixing and changing some codes but we have literately heard zero from Wildcard about this issue. Nvidia tried to fix it with the new drivers but did not work so wildcard need to set this a a top priority on this matter. 

More people will jump to 2000 series more people cant play, if thats not game breaking..

And fyi, i have tried the game on 2 different PCs with 2 completely different hardware.  1 with 2080ti and 1 with 2080, i7 4970 and 8086K new ram different PSU everything (not SLI). Same issue same kernel log.

Its not hardware related at this point.

[CyberAngel67]

44 minutes ago, iStephano said:

I can agree on that GeForce Experience can create issues. Have seen that many times.

But we need to look at the logic here, 100s of people have the same problem and even get the same kernel crash log. I mean if 2 people got that sure it could be a hardware issue with their PC`s they are using. But we are talking about a lot of people here with different setup that gets the same exact problem. We have all tested the game the same way, like jumping on a flyer on the island map. 5 mins later BSOD.

This problem is in Ark itself as it could be that the game is outdated and can not run 100% on the new card. New architecture and of course Nvidia may have failed at some point. But as i have seen and read this is a software issue.

And dont give me that PSU nonsense i have a 1000w PSU and its giving more than enough power to all components in my PC.

Every other game i have tested on RTX 2080 to this date runs fine for days, for weeks. No windows issues, no game crash, no BSOD, nothing it runs just fine. but as soon i load up the island map with the RTX 2080, boom BSOD 5 mins later. And its always when the game itself load new textures or to be more specific jump on a flyer so the map needs to load stuff in faster.

The new DLC seems to run fine so maybe they fixed something here, but they need to apply that to the older maps as-well.

This is a software issue.

You run on older card with a older architecture thats match with Arks game code better. So whatever you says is invalid sorry to say it. With older cards the driver also works better as you can rollback to a more stable one if something is bugging here. I dont know maybe you play Ark 20mins a day? Some people can spend 3-4h a day and thats when the issue on older cards kicks in.

And another thing, if a person buys a game you can not expect that person to go into BIOS setting to "fix things". It does not works this way. if that was the case most people would stick to console gaming and throw pc gaming out the window.

I fully understand the problem with new hardware and "older games" can crash, its needs some fixing and changing some codes but we have literately heard zero from Wildcard about this issue. Nvidia tried to fix it with the new drivers but did not work so wildcard need to set this a a top priority on this matter. 

More people will jump to 2000 series more people cant play, if thats not game breaking..

And fyi, i have tried the game on 2 different PCs with 2 completely different hardware.  1 with 2080ti and 1 with 2080, i7 4970 and 8086K new ram different PSU everything (not SLI). Same issue same kernel log.

Its not hardware related at this point.

Well I get this BSOD on GTX 1080 TI's as well.

Always the same, Device lost, close the crash box boom BSOD.

[RageQuitter]

24 minutes ago, CyberAngel67 said:

Well I get this BSOD on GTX 1080 TI's as well.

Always the same, Device lost, close the crash box boom BSOD.

Install 399.24, the problem is the 4xx.xx drivers, not had one BSOD since I changed

[dokany]

AMD R7 2700X

TUF X470-PLUS

RTX2080

My problem is when I entering the surface on aberration, I got BSOD. Otherwise, that's fine....Everywhere is ok but surface.

So I can only use another video card to enter surface.

[CyberAngel67]

2 hours ago, RageQuitter said:

Install 399.24, the problem is the 4xx.xx drivers, not had one BSOD since I changed

doesn't fix the issue

[DrunkMunki]

8 hours ago, BluesDS said:

I'd also like to add its been shown this doesnt seem to happen with two cards in SLI. I've been lurking this thread for weeks and this comment finally pushed me to speak

why are you using SLI when its no longer supported?

 

I just got a 2080 and had no issues, i had a 1070 which i was using (from my VR rig) and it was crashing a few mins in....

[RageQuitter]

4 hours ago, CyberAngel67 said:

doesn't fix the issue

Check the driver version, I left the "install geforce experience" box checked first time and it installed the latest driver without asking, ended up removing anything to do with nvidia then installed it again and unchecked the box

It's definitely the 4xx drivers causing it, I couldn't play for more than a couple of hours without it happening then changed drivers nearly 3 weeks ago and not had one BSOD since

[DarthaNyan]

16 hours ago, Xenithar said:

Also, for those stating you need drivers below 400.xx, stop spreading false rumors. As I have stated for ages, I am on 416.34 and have been for a while now. Two 1070's in SLI. Zero crashes since disabling the RHI thread. I have played all day a time or two on the weekends. No issues. I even have TrueSky enabled. Blaming the drivers is just plain wrong.

IF you dont have a problem - that doesnt mean others dont have it. Also i remember @Chris saying that he runs 2080Ti in Sli with no problems/crashes as well.

So maybe the solution to the RTX BSODs is to buy more RTX cards?

[CyberAngel67]

4 hours ago, RageQuitter said:

Check the driver version, I left the "install geforce experience" box checked first time and it installed the latest driver without asking, ended up removing anything to do with nvidia then installed it again and unchecked the box

It's definitely the 4xx drivers causing it, I couldn't play for more than a couple of hours without it happening then changed drivers nearly 3 weeks ago and not had one BSOD since

It is 100% 399.24, I am aware how to make GeForce keep that version. In my case I have had my GTX 1080 ti for 12 months, in that time it has crashed with Device lost 2 times and more recently 3 times with a BSOD.

Now while everyone used to get the Device Lost issue, I wasn't actually getting the crash on my 1080 TI, but they where being reported in the Event Viewer.

And the last BSOD was 100% 399.24 driver version installed.

[XiangZhu]

23 hours ago, XiangZhu said:

I'll probably do more testing with -norhithread and reduced graphics settings this week - see if I can get a temporary workaround to the problem for myself or identify exactly what to turn off to make it all better in the short run.

Ran with the -norhithread argument for an extended period last night: Received one BSOD returning to base for like the 4th or 5th time. Thought it was a-ok, but no. Switched to DX10 for a bit. Guess that will be my recourse when I need to get some stuff done for a while. Sigh.

[XFROGGYMANX]

22 hours ago, Xenithar said:

Actually, GeForce Experience attempts to configure settings for you in games and does other dopy things, like updating drivers automatically. You know, drivers which are working FINE as-is. It causes way more problems than it solves. I'll control my drivers and settings, thank you.

Also, for those stating you need drivers below 400.xx, stop spreading false rumors. As I have stated for ages, I am on 416.34 and have been for a while now. Two 1070's in SLI. Zero crashes since disabling the RHI thread. I have played all day a time or two on the weekends. No issues. I even have TrueSky enabled. Blaming the drivers is just plain wrong.

Now for a lesson in BSODs. Did you know that a BSOD is triggered by a hardware error? That's right, kids! No game can tell Windows to blue-screen! This takes a low-level driver which has detected an error, such as low voltage, high voltage, or some other physical error somewhere on your system. This is why I recommended checking the power supply. A single 450W PSU may not be enough to run that shiny new 2080 in 4k with max details. My rig required a 600-650W PSU, so I bought a 750W PSU. Low voltage can cause a BSOD just as fast as an overheated CPU or GPU.

Again, drivers above 400 are fine. If not, please provide solid evidence and don't just throw around a theory you have. Plenty of us are running above 400 with 1070's, 1080's and even the 2000 series without a problem. If the drivers above 400 cause an issue on your system, then you need to figure that out first.

did you even take the time to read the thred sli cards or duel cards have been stable this is about single 2080s and 2080ti also have a 1200 psu and my rtx just sits in a box collecting dust because every time i load in a map and hop on my flyer i get a BSOD BAD-POOL-CALER and d3d11 fatle error in game only happens in ark and on maps with high LOD that being said my 1080 works fine no problems at all 

 

[Anarki]

I had this issue too and so many people have said so many things that in my opiion - noone knows how to currently fix this.

Im using a i7 7800k CPU

Win 10 Pro

Previously a R9 295x2 (but was getting D3D device hung/lost error) and am now using a 1080ti

1200i Corsair PSU

32gb DDR4 Corsair Vengeance Pro RAM

Had the issue on Center every 2 mins on old card, now with the 1080ti it happens once in playing session but no error message - just crashes. So its happening a lot less. People say its Unreal engine, others say its Win 10.

Can we all possibly try to identify a common factor here? Put your specs down maybe?

It did seem to help running with DX10 rather than DX11 on my old card if that helps anyone.

[Xenithar]

22 hours ago, XiangZhu said:

Obviously, if you have to throw in the -norhithread, then you have an issue. Please don't say that you don't. 

Either the devs rewrote the rhi classes of the unreal engine, which I highly doubt, or nvidia made a boo boo and they need to fix it. If they didn't make a boo boo they need to do a work around to fix the 'natural' issue that's occurring since the new drivers came out for the RX cards.

I'm glad the -norhithread argument is working for you, but I'd ask that you not talk down to the rest of us in this thread.

Actually I don't. I would suggest a good read over the forums at Epic. RHI is something that was designed to work with DX12 and Windows 10. I run 7 Pro 64bit, so no DX12 and no 10. I dislike 10 on many levels and DX12 is currently a mess, so I am sticking with 7 for the time being. It will be supported through 2023, so I am good.

Also, the RHI thread is supposed to be separate on consoles and DX12 systems ONLY. This means those of us on 7 or prior must use the command. For whatever reason, Ark does not check the DX version and not use the thread if on 10 or 11. Questions about this have been asked, but not answered.

Also, the Unreal Engine, prior to 4.16 (we are using a modified 4.5 unless I am mistaken) cannot use the shader cache unless you pass "-norhithread" to it upon startup. They actually note this in the patch notes for 4.16! It is down the page a bit. Maybe 3/4 of the way down in the "Release Notes" section. Perhaps shadowing on these new cards uses said cache for whatever reason.

Also with the D3D device lost issue that Ark has struggled with for years, the "-norhithread" parameter was actually in patch notes! This has been discussed plenty of times. Heck, look at the second reply here!

I can go on and on about this, providing thousands of discussions on the point, but the rule of thumb is that the RHI thread is not the best in the world on many systems and configurations. Heck, the AMD sub-reddit has discussions about UE4 and this very thread. I run Intel so I cannot comment as to whether or not AMD is more likely to have issues. Nothing against AMD, I just prefer Intels offerings.

17 hours ago, iStephano said:

I can agree on that GeForce Experience can create issues. Have seen that many times.

But we need to look at the logic here, 100s of people have the same problem and even get the same kernel crash log. I mean if 2 people got that sure it could be a hardware issue with their PC`s they are using. But we are talking about a lot of people here with different setup that gets the same exact problem. We have all tested the game the same way, like jumping on a flyer on the island map. 5 mins later BSOD.

This problem is in Ark itself as it could be that the game is outdated and can not run 100% on the new card. New architecture and of course Nvidia may have failed at some point. But as i have seen and read this is a software issue.

And dont give me that PSU nonsense i have a 1000w PSU and its giving more than enough power to all components in my PC.

Every other game i have tested on RTX 2080 to this date runs fine for days, for weeks. No windows issues, no game crash, no BSOD, nothing it runs just fine. but as soon i load up the island map with the RTX 2080, boom BSOD 5 mins later. And its always when the game itself load new textures or to be more specific jump on a flyer so the map needs to load stuff in faster.

The new DLC seems to run fine so maybe they fixed something here, but they need to apply that to the older maps as-well.

This is a software issue.

You run on older card with a older architecture thats match with Arks game code better. So whatever you says is invalid sorry to say it. With older cards the driver also works better as you can rollback to a more stable one if something is bugging here. I dont know maybe you play Ark 20mins a day? Some people can spend 3-4h a day and thats when the issue on older cards kicks in.

And another thing, if a person buys a game you can not expect that person to go into BIOS setting to "fix things". It does not works this way. if that was the case most people would stick to console gaming and throw pc gaming out the window.

I fully understand the problem with new hardware and "older games" can crash, its needs some fixing and changing some codes but we have literately heard zero from Wildcard about this issue. Nvidia tried to fix it with the new drivers but did not work so wildcard need to set this a a top priority on this matter. 

More people will jump to 2000 series more people cant play, if thats not game breaking..

And fyi, i have tried the game on 2 different PCs with 2 completely different hardware.  1 with 2080ti and 1 with 2080, i7 4970 and 8086K new ram different PSU everything (not SLI). Same issue same kernel log.

Its not hardware related at this point.

You need to study computers a bit before inserting your foot into your mouth. First, "every other game you tested" likely isn't running UE 4.5, so the test is invalid. May as well compare a Honda Civic to a big-rig while we're at it. Again, software cannot cause a BSOD. Hardware causes that. Learn Windows, then comment.

As for the "older card", you're right. It didn't come out today, so it's old by your standard. Good going! Guess the brand-new 2018 cars are "old" despite being just made, because a 2019 rolled onto the lot also, right? And if you REALLY want to get down to it, I ran Ark for a year on an i5-2500, 16GB DDR3, 500GB SATA2 disk, and a whopping GeForce GTX 550 Ti with 1GB of VRAM. Ran fine on medium in 1600x1200. I also suggest you check the specs required when the game was released. I believe the minimum requirement was a GTX 670. I was well below it and had no issues. So before we claim that the second most-powerful card from the generation that was only replaced a month or so back is "old", we need to get our facts straight. After all, if a 550 can run it fine, a 670, or even a 1070 can, then why can't your shiny new card handle it. Did you even try disabling the RHI thread as suggested? You never mentioned that.

Oh, and you should read my previous post. I played all day several weekends. That is 8+ hours. No crashes. If you would read before commenting we could have skipped this part.

Your BIOS comment is just stupid. Let me get this right. You are so simple you cannot press F2, DEL, or whatever on startup and use your mouse or keyboard to uncheck a box, but you find it simpler to go into your registry and try to change hexadecimal values which can break Windows, in an attempt to override a sleep-state which is easily disabled in the BIOS? Really? Care to explain this? It's this mentality that keeps systems coming to my shop for repair, by the way. Keep messing around in the registry. Keep screwing things up so I can get paid.

Finally, you destroyed your own last point. You claim it isn't hardware-related but you're running a 2080 in BOTH systems. That's like buying a Dodge Neon versus a PT Cruiser and telling me it's a different drivetrain. Same engine and transmission there, homey. Since this is a graphics-related issue for you, the hardware in both systems related to graphics rendering is the same.

 

As for the PSU debate, you DO have a more than adequate PSU, as do some of the other posters here. However, many do not. For example, "dokany" mentions owning an AMD R7 2700X with a 470W PSU. Assuming he has a basic CD/DVD drive, one SATA disk, and 16GB of RAM in quad-channel, he needs a MINIMUM of 574W PSU. The duty-cycle will call for about 520W, which is 50W over the max his PSU can spike to. I will link to a fairly accurate PSU calculator below. Maybe it can help some of you. Those of you who are running 1000W or more and NOT running 8-disk RAID coupled with quad SLI, you're fine and dandy, whether you run Intel or AMD.

The point here is that just because YOUR PSU is fine, don't speak for everybody. Low voltage easily triggers blue-screens. Get an adjustable circuit, plug your PSU into it, boot Windows and start lowering the voltage. Eventually you will have a BSOD due to something not being powered correctly.

PSU calculator

*EDIT*

I'll take it one step farther. I work in the IT and IS fields. I debug crash dumps from time to time to resolve corporate issues. If you can make your system BSOD and send me a copy of your crash dump, I'll take a look at it. Not the Ark dump, the Windows dump. I can't do much with an Ark crash stack or dump.