Question RTX 2070 seems fine but is causing system crashes ?

[Lemonman74]

In certain games (armored core 6, elden ring and other modern titles) the 2070 mini will cause the pc to crash with a black screen and fans at 100%. The issue also happened when i put it into another system. There seems to be no power/ thermal issue as i can stress test the card and get it to work at 100% usage for ages with the card not causing any issues and not getting hotter than 76 degrees.

I tried cleaning the card and adding new thermal paste, updating the drivers, rolling back to older drivers, underclocking and flashing the vbios. The issue even persisted on a linux partition. I have an older r9 390 which consumes a lot more power and it works without issue.

Any idea what this issue could be? The card works flawlessly otherwise and the pc is entirely usable unless you launch a game it doesn’t like.

 

[Lutfij]

Welcome to the forums, newcomer!

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

updating the drivers,
Using DDU? Manually reinstalling the latest GPU drive in an elevated command?

 

[Lemonman74]

Welcome to the forums, newcomer!

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

updating the drivers,
Using DDU? Manually reinstalling the latest GPU drive in an elevated command?

Hello and thanks for your reply.

I installed the drivers using DDU in safemode and installed new drivers in elevated command prompt.
I also tried installing using nvcleanstall but that did not resolve the problem.
I also tried disabling xmp and loading optimized defaults in the bios.

On one occasion after installing the latest nvidia drivers the pc would boot and would be unable to load windows properly needing a restart until it began automatic recovery. I had to launch into safe mode to uninstall the drivers to fix this.

I'm on the latest drivers now and i can't get that behavior to reappear so it may be a fluke but it makes me think this is a simple software issue.

I got this graphics card from a friend who bought it used and had the same issue. The card even has an "OK" green sticker on it from a repair shop somewhere, they must have stress tested it and it ran fine.


Here are my specs:

CPU: R5 3600
CPU cooler: Noctua NH-U12s Redux
Motherboard: MSI b450 tomohawk max w/ bios ver. 7C02v3I (latest)
Ram: 16gb corsair vengeance
SSD/HDD: samsung 850 evo with a 2ttb seagate hdd
GPU: Rtx 2070 mini 8gb 1 click OC driver version 551.76 (latest)
PSU: evga 650w gold rated/ about 5-6 years old as i said never any issues even with more power hungry cards
OS: Windows 10 currently, issue also occurred on a updated linux mint install
Monitor: Acer R241Y

 

[Lemonman74]

Update:

I've done a lot of testing and I can say I'm mostly confident this is a driver issue.

Two games which would crash quickly and consistently were armored core VI which would crash shortly after the main menu and The Finals which would crash in the main menu or seconds after entering the practice range which I used as a benchmark.

The black screen of death would occur on both of these games with the latest drivers (551.76), the second latest drivers and the oldest I could find (416.34).

However, when I launched the finals with the older drivers I got a warning message saying something along the lines of "There are known issues with your current driver. Please update drivers or don't use the d3d11 api recommended version 545.84".

After installing driver version 545.84 I have tested both games. The Finals ran for 32 minutes of gameplay without crashing at all. I tested Armored core VI which got passed the main menu and was able to load a level which I tested for 15 minutes without issue. I have not tested Elden ring as I don't have it installed but I would assume it works also since the other two were fine. I used the same install method with all drivers only 545.84 seems to work.

I haven't got it to black screen on this specific driver yet. I will reply if it does eventually but I think its fixed.

A driver issue with the graphics api explains why only some games were crashing and some games were fine.

Have you ever heard of issues being fixed with driver updates then being reintroduced later down the line?

I've seen a lot of black screen of death issues on forums like these specifically with the RTX 2070 and RTX 2080 so I hope this helps any of those people.

 

[Mahisse]

One cause of crashes like this is bad memory. It's often that people neglect on this first time around and focus too much on the GPU. Especially if it's not running hot and it doesn't seems like other users have issues with drivers. It's a tedious job to test bad memory sticks but it can save you headaches in the end (ie buying a new GPU and the problem isn't solved).
It's tedious depending on how many meomory sticks you have because you need to test each stick individually to test correctly.. and gets even worse if you suspect it can be a bad memory slot on the board.

Anyhow, to me it sounds like it's very likely could be bad memory because crashes happens randomly. That makes sense due to the nature of how RAM is working. For games not using much memory you may never run into the faulty section of the memory. The higher memory uses the more often crashes happens.

Have you ever heard of issues being fixed with driver updates then being reintroduced later down the line?

I'm sure that has happened in the past but it's more likely the problem is on your end. Perhaps you also want to check for BIOS update.

 

[Lemonman74]

Thanks for your reply.

The crashing issue occurred on multiple systems which worked well with prior graphics cards. I could not recreate the crashes even running the same games with a different graphics card. I'm certain the 2070 was the culprit. I have not had it crash on me yet when previously I could reliably crash the pc in seconds which I used to test the drivers.

Just in case I ran windows memory diagnostic and no errors were found.
I know the solution sounds like it shouldn't work but installing a specific older driver appears to have fixed it. I have no idea why but it works.

Thankyou for your suggestion anyway as I have had faulty memory on a system causing bluescreens a few years ago.