Hi all- 
Looking for some advice to trouble shoot a problem that I've been running into the last week. 
RTX 3090 FTW3 Ultra
 - Replaced in March 2021 via the 3090OC Special RMA Program 
 - Replaced via RMA due to PCI Slot pulling out of spec Wattage (too high)

I've been experiencing Blue Screen system crashes when playing Amazon New World. I do know some others have had problems with this game but this is a new phenomenon for me. I've tried to include the relevant details below and happy to include additional info if folks think that will help.
 - Issue began in the last 3-5 Days.
 - I've played New World hundreds of hours using identical setting for months. I'm very anal about software and driver updates; so everything is 100% up-to-date.
 - I undervolted my card at 850mV and at 1830 MHz via afterburner.
    --This is the same under volt I have used across multiple games including this one for months and normaly everything is rock solid.
 - I'm now getting periodic Blue Screen crashes; I thought it may be due to MSI Afterburner so decided to uninstall and just use PX1...this is when things really got weird.

• If I attempt to play New World using 100% Stock setting in PX1 system will blue screen crash 100% of the time. Sometimes this takes 30secconds, sometimes 5-10min...the but regardless even on stock settings I am getting these blue screens.
• Tried re-installing drivers and PX1 and did not resolve the problem. 
• Tried running some other games like Witcher 3 and AC Valhalla and everything seems ok.
• I even ran some Benchmarks with PR and Timespy and rock solid. Did some light OC-ing and was able to get PR scores around 14,200 (this is using the Normal not OC switch). Basically everything seems normal. My RMA card has always run like a dream prior to now. On cold winter days I was able to top out PR scores around 14,900 on stock Air.

 
Currently I am able to trigger the BSOD crash by Playing New World UNLESS I'm using MSI AB to lock in a under volt. 
 - Note I do play under volted at 850mV 95% of the time but sometimes used slightly more aggressive under volts at 900mV or 950mV...with a slightly higher boost MHz- especially on cold days. I almost always keep the card under 60/65 and if I stick with the 850mV undervolt it hardly ever tops 60 Celsius in the winter. None of this ever gave me even the slightest problem as I always error on the side of a conservative Overclock. 
 - Now, with this problem, anything but my most conservative UV seems to trigger the Blue Screen system crash. IN fact, I am getting crashes even when using the UV settings I've used for 500+ hours of gameplay...though these are much less common if I play for a few hours I will 100% trigger a blue screen crash. And, as noted, if I try to play New World with 100% Stock settings or anything other than a very conservative Undervolt the game will crash within seconds or a few min. 

• If I play with stock settings and power locked to 100% the card with normally fluctuate around 900-1020 mV and somewhere in the 1870-1960 MHz range. This should never cause blue screen crashing since it is literally stock and the temps are under 70 Celsius. However, the crashing persists and the only way to actually get the game stable to to under volt (and its still not 100% stable, but at least i can run it for a couple hours normally this way).
• This was never the case before; I prefer the stability and lower temps of a UV so play that way 95% of the time. However, if I did play New World with the power limit cranked to 107% or using the OC switch...I never had blue screen crashing...just ran a bit hot and noisy- but still stable always. This has been the case all the way back to the beta and launch 6 months ago...now things are oddly unstable.
• I've tried some basic trouble shooting like a full game re-install, MSI and PX1 re-install (I never run both programs and once and totally uninstalled MSI while troubleshooting and crashes continued).

 
Other than isolating this as a Video Card issue (seemingly linked to a single game) I'm stumped and cannot imagine what caused such a drastic drop in stability using the same settings and practices that have worked perfectly for months. 

Looking for any helps or ideas. Trying to avoid a full system wipe. Thanks in advance.
 

Sounds like the card has degraded. No need to do a system wipe when you can use windows system restore to restore your computer to a day when everything was working correctly.

As you suggested I tried setting the Power Target in EVGA tool to 70%; I also tried at 80%. Both times I was able to run New World for about 45min+ with no issues. I'm continuing to test this but that does seem to somewhat address the issue as when I keep the power limit and all setting stock I get the blue screen in under 2min. 

I'm going to do some additional testing in other games where I apply minor overclocks to try and determine if this is isolated to just New World. I did some light testing in Witcher 3 with a minor overclock and power limit set to 107% and things seemed fine. If the card is degrading I would want to RMA it...though that would suck as I feel I got really solid silicone on this one- or well, up until now I thought that as I was able to get PR Scores near 15,000 using stock air cooling.

Thank you for your help. I will continue testing today and see what happens with a longer test in New World and some 1hour+ testing in other games with a Overclock applied. 

What bug check codes are you seeing?

AndrewGSQ

Pugwoof
Checking in here again as I unfortunately think it is time to RMA....again. 


I'm getting blue screen crashing when running a bunch of the OC Profiles I had setup 6+ months ago when playing games like Witcher 3 or AC Valhalla. These same OC were rock solid for hundreds of hours of play back in early to mid 2021. 

So far as New World...the game blue screens even if the card is running 100% stock. If I undervolted to 850mV using MSI AB it is mostly solid- but even this has started to give an occasional crash. I was using this and more aggressive UVs for 500+ hours of play between Summer 21' Beta and December of 21'. If I use EVGA PX1 and lock the power to 80% I do not get the blue screens...but from my position if I cannot run the card with stock 100% power in ANY game (even a poorly designed one like New World)- that is an inexcusable problem and something is wrong with the card. 
--I'm especially concerned since a couple months back I could run conservative Overclocks in even New World and did not get blue screens...I just prefer Under volting for lower Temps and fan noise.
--Something has definitely changed (gone wrong) in the past 4-6 weeks...perhaps the card is degrading? 

Wanted to get some input here from folks on the forums. Is there anything worth trying here to trouble shoot? Or do I just suck it up and RMA. Would be a shame since I did seem to do pretty damn well with the Silicone Lottery on this card. Last March/April when I was benchmarking the card after my 1st RMA I was pushing close to 15,000 on Port Royal on stock Air Cooling.
--I plan to call EVGA Support Monday when they are open again.
--Would appreciate any advice or ideas on possible fixes I could try in the next couple days before I go for the RMA. 

Thanks in advance!

Can you just return it for a new one or is that not possible because of the current situation?

I have 700 days left in warranty...so should be able to RMA no problem. Was just seeing if there was any way to avoid this and perhaps something I could fix- but not looking that way. 

I've not been home for 4 days so did not get a chance to record what the Blue screen says. Once I get back tomorrow I will provide info on the error that appears when blue screen occurs.

I have a 1080ti before with default settings underload will crash the programs the only way for it to work was under volt, I ended up RMA the card.

I'm going to pick up a PSR and try this out. Sorry for not responding. I bounce between two locations and PC is only at one of them so could not really check anything. 

I have a EVGA G3 1000 watt PSU; I'll pick up another EVGA (as I have custom cables so need to stay in brand)...I'm hoping that is the problem as PSU is also in warranty and MUCH easier replacement via RMA. 

OK so I picked up a EVGA 1000 GT (Gold Rated PSU) at Microcenter...I don't think it is quite as good as my 1000 G3; but should work to determine if the issue is the PSU or the GPU. 

I also confirmed that my custom cables are cross compatible so no need to drive my self insane rewiring everything- since I have my wires is a very particular way for aesthetics.
 
My plan is to swap out the PSU and run a bunch of scenarios where it was previously crashing. I took off Thursday from Work so hopefully will have time to get to this before end of the week. 
 - I figured also a good idea to dust/clean the whole system a bit while I'm swapping out the PSU. I already keep things very clean but can never hurt to pop out the GPU and blast out any dust my regular cleaning may miss. 
 
My thinking is if running the card stock in games like New World, or with very modest OCs in games like Witcher 3 lead to instant BSOD (as it does now)- GPU needs an immediate RMA. If the BSOD cease in these same situations- 'G3' PSU needs a immediate RMA. Here's to hoping this is the PSU and I'm not going to be replacing my 2nd 3090 FTW3 Ultra in 11 months. 

My thoughts- it's your custom cables that are unstable / wrong/ worn
 
next thought would be to OVERvolt & see if that does the trick ....the 30xx series cards are very finicky of power, between the firmware updates & the nvidia driver fixes , your current card may need extra juice now , not less .
 
didn't see you mention if you're using a pcie riser cable ? - but that should also be removed from the equation if you do ...

Pugwoof

aka_STEVE_b
My thoughts- it's your custom cables that are unstable / wrong/ worn
 
next thought would be to OVERvolt & see if that does the trick ....the 30xx series cards are very finicky of power, between the firmware updates & the nvidia driver fixes , your current card may need extra juice now , not less .
 
didn't see you mention if you're using a pcie riser cable ? - but that should also be removed from the equation if you do ...

Not using a PSI Riser cable. And these are the same Cables I used for 10 months without any issues whatsoever- they were brand new then...meaning they are not even 1 year old. I used these same cables with no changes to run games and benchmarks for hours with solid Overclocks or even stock and never had a problem before these BSOD started 6/7 weeks ago. 

As noted in previous posts any overvolting results in immediate BSOD when launching a game; same as with even mild Overclocking. SOME games run just fine at stock, others like New World will crash even with 100% stock settings. If I undervolt or cap the Max power with slider at 80/85% then New World is stable. I cannot get anything to run stable with even a very very Modest OC...like +20 Core / +90 Memory. And, as noted, some games even BSOD at 100% stock. 

I can try to test with the cables the new GT 1000 PSU came with I suppose...seems unlikely that cables are the issue as they've not changed.

I wouldn’t rule out the power cables. I had a SATA power cable that died in 2020. It was part of a CableMod kit and was honestly the last thing I suspected. It worked fine for about 7 months before crapping out. Thankfully, CableMod sent a replacement.

ObscureEmpyre

Pugwoof

aka_STEVE_b
My thoughts- it's your custom cables that are unstable / wrong/ worn
 
next thought would be to OVERvolt & see if that does the trick ....the 30xx series cards are very finicky of power, between the firmware updates & the nvidia driver fixes , your current card may need extra juice now , not less .
 
didn't see you mention if you're using a pcie riser cable ? - but that should also be removed from the equation if you do ...

Not using a PSI Riser cable. And these are the same Cables I used for 10 months without any issues whatsoever- they were brand new then...meaning they are not even 1 year old. I used these same cables with no changes to run games and benchmarks for hours with solid Overclocks or even stock and never had a problem before these BSOD started 6/7 weeks ago. 

As noted in previous posts any overvolting results in immediate BSOD when launching a game; same as with even mild Overclocking. SOME games run just fine at stock, others like New World will crash even with 100% stock settings. If I undervolt or cap the Max power with slider at 80/85% then New World is stable. I cannot get anything to run stable with even a very very Modest OC...like +20 Core / +90 Memory. And, as noted, some games even BSOD at 100% stock. 

I can try to test with the cables the new GT 1000 PSU came with I suppose...seems unlikely that cables are the issue as they've not changed.

I wouldn’t rule out the power cables. I had a SATA power cable that died in 2020. It was part of a CableMod kit and was honestly the last thing I suspected. It worked fine for about 7 months before crapping out. Thankfully, CableMod sent a replacement.

I cannot believe it but it seems like the cables are likely the problem here. Sorry for the long time between responses as I took a vacation and only just got back. 

Anyway, I swapped the custom CableMod 8 pins for some generics that came with the EVGA 1000 GT PSU I bought to help process of elimination testing (as noted above). To my surprise this seems to have fixed the problem. When "aka_STEVE_b" mentioned this above to be honest - thought it was dumb and no possible chance this was the problem. Happy to say looks like I was wrong! 

I'm still only early in my testing and leaving for 48 hours again later today...but I plan to do a bunch more testing this week as I need to return the PSU I bought from Micro Center by Friday. I was able to keep all my normal Undervolts and some modest OCs stable for a couple hours- these were previously crashing in mere seconds or at best a couple min. I was able to play New World and Witcher 3 with a overclock for a couple hours without any problems. Before the switch ANY overclock would instant crash and New World would crash on 100% stock settings within 30sec.

I will say my PC looks like crap right now- but if it works and I avoid a 3090 RMA who cares! I don't want to get too excited until I can really do a full 6-8hours of in game (and benchmark) testing to really ensure things are just as stable as they were prior to this problem.


Wish me luck that this is indeed only a cable problem and I can get this closed out this week. I appreciate all the help and replies over the last month. As for the most recent question I am on the latest Nvidia drivers.