• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Need Advice – Faulty RTX 3070 Ti Crashing Under Load

Associate
Joined
9 Apr 2025
Posts
3
Location
East of England
Hi everyone,

I’m hoping someone here might be able to offer some insight or advice. I’m not hugely tech-savvy, so please forgive me if I explain anything poorly. I’ve been dealing with an increasingly frustrating issue on my system, and I’m reaching the point where I don’t know if there’s anything left to try, or if I should just sell the GPU as faulty and cut my losses.
I have a Gigabyte RTX 3070 Ti, It boots fine, outputs display, and works on the desktop with no issue at all. The problem is that whenever the card is under any sort of GPU load, it either crashes the game back to desktop, freezes the application for several seconds, or causes a display driver timeout, worst case has been a total freeze of the computer, to which I had to hard-reset, though this has been a very rare occurrence.

This happens in pretty much every game I’ve tested and even a few testing environments such as Kombuster, MSI Afterburner, etc

What I’ve Done So Far:​

  • Multiple clean driver installs using DDU in safe mode
  • Tested a wide range of driver versions (from 522.25 to the latest 572.x)
  • Undervolted and underclocked the GPU via MSI Afterburner (down to 1800MHz / 900mV and lower)
  • Tried multiple power limits and fan curves
  • Reseated the card and cleaned contacts with compressed air
  • Swapped PCIe slots (note: card only works in the top slot — nothing at all in the second)
  • Disabled PCIe Gen4 and Gen3, forced Gen2 in BIOS
  • Fully reinstalled Windows 11
  • Logged detailed telemetry with HWInfo (no abnormal temps, voltages, or power draw before the crash — temps stay below 50°C)
  • Tried limiting FPS and lowering game graphics settings
  • PSU is a Corsair 1000W Gold modular, and everything else in the system is stable
The Errors I See:
  • Event ID 13 – “Graphics SM Warp Exception: Illegal Instruction Encoding”
  • Event ID 153 – “Reset TDR occurred on GPUID:100”
These only happen during GPU load. The system doesn’t bluescreen, and the rest of Windows works fine after the driver crashes — it’s purely a GPU rendering issue.
I’m just trying to decide if it’s worth sending off for repair, or if I should just list it on eBay as “Faulty / For Parts” and be done with it.

Apologies if this is long-winded or if I’ve missed anything obvious — I’m not a hardware expert, just someone trying to make the most out of what I’ve got. Any advice, suggestions, or direction would be massively appreciated.

Thanks in advance.
 
I’m just trying to decide if it’s worth sending off for repair, or if I should just list it on eBay as “Faulty / For Parts” and be done with it.
If the card is under warranty then I'd get it sent back.

The problem is that whenever the card is under any sort of GPU load, it either crashes the game back to desktop, freezes the application for several seconds, or causes a display driver timeout, worst case has been a total freeze of the computer, to which I had to hard-reset, though this has been a very rare occurrence.
I assume:
1. The card previously worked?
2. You have tested with another graphics card and/or integrated graphics?
 
If the card is under warranty then I'd get it sent back.


I assume:
1. The card previously worked?
2. You have tested with another graphics card and/or integrated graphics?
Already tried reaching out to Gigabyte about the issue but because the cards warranty is up, they basically refused to help.

Card worked up until recently, Installed the latest Nvidia driver (527.70) and all of a sudden everything went wrong. As I said above, I've used DDU several times and reverted back to previous version, even the version I was on prior to updating but its just not working.

Yeah, I've tested the card in several other rigs and each time, they display the same issues so, it's 100% the card.
 
Card worked up until recently, Installed the latest Nvidia driver (527.70) and all of a sudden everything went wrong.
Yikes, unfortunately you're not the only one.

Already tried reaching out to Gigabyte about the issue but because the cards warranty is up, they basically refused to help.
There are a fair few repairers out there, but obviously some are better than others. This guy I watch semi-regularly:
 
Yikes, unfortunately you're not the only one.


There are a fair few repairers out there, but obviously some are better than others. This guy I watch semi-regularly:
Yeah, I've been hearing that a lot of people have had issues with the last few driver updates. Thats why I originally thought it was a driver issue myself.. but this is why I feel like its either VRAM controller or memory interface issues. I just am not entirely sure because this isn't my area :cry:
 
Yikes, unfortunately you're not the only one.


There are a fair few repairers out there, but obviously some are better than others. This guy I watch semi-regularly:
hello internet! this is a guy that will actually fix your gpu if possible not like another clown who runs a repair shop and if it takes too much time its a no fix and on to the next
 
I just am not entirely sure because this isn't my area :cry:
From what I've watched with the guy I linked, even if you know your stuff, it is near impossible to diagnose without having a proper look at it and even then, there's a degree of trial and error.

There have been some cards that do that, just crash/freeze when starting a game and from what I can recall, the fixes are anywhere from very simple to rather complex (mostly with the troubleshooting). The fact the card works entirely fine on the desktop and there's no visual corruption suggests to me that the fix is likely to be on the simple side, but yeah, it is way out of my area too.
 
Sounds like the card is overheating, consider a re-paste of the GPU chip. Then test with Furmark.

If Furmark shows anomalies while running, you have failing memory chips.
 
hello internet! this is a guy that will actually fix your gpu if possible not like another clown who runs a repair shop and if it takes too much time its a no fix and on to the next
TBH there's a few industries like this, esp some car places or they firing the parts cannon. Unless theres someone in the UK that does repairs by the time you send the card to the states and pay for the diagnostics and repair and get it shipped back to you this isnt going to be cheep and with most cards the cost of this one might as well get another card.
 
I have had the same issue for a few months starting with my rtx3060 - after it was running OK for a couple of years. When it crashed under load it would bluescreen me and I would have to open up the case and flash the bios to get the PC to reboot. I had the card changed to an RTX4060 under warranty (thank you Overclockers!) but this just led to the game crashing - PC not dying now, but I still can't play any newish games, even on lowest settings. I tried all the online tricks by underclocking, reinstalling, monitoring resources, scanning everything, reseating everything, repairing everything. Nothing has worked.

The PC has been back for repair twice and to be honest I've lost patience and ordered a new one.

But if anyone ever finds a fix for this I wouldn't mind repairing this PC in case my grandson wants it.
 
I had the card changed to an RTX4060 under warranty (thank you Overclockers!) but this just led to the game crashing - PC not dying now, but I still can't play any newish games, even on lowest settings. I tried all the online tricks by underclocking, reinstalling, monitoring resources, scanning everything, reseating everything, repairing everything. Nothing has worked.

But if anyone ever finds a fix for this I wouldn't mind repairing this PC in case my grandson wants it.
By "crashing", what exactly do you mean? CTD? BSOD?

You have reinstalled Windows already, I assume? You're not using the latest nvidia drivers, right? They have been widely reported as being bad.

Have you run memtest to check your memory is good?

If everything has been running fine for a few years I'd check your temps are good and given what you've tried already, you're close to the point of replacing parts.

By the way, was the 4060 a b-grade, or a new one?
 
Last edited:
What is your Gigabyte RTX 3070 Ti model?

What is your motherboard brand and model?

What is GPU temp in games?

How old is Corsair 1000W Gold modular PSU?

Do you have any AMD software installed? Then uninstalled all of AMD software.

Try update both motherboard and GPU to latest BIOS version.

Try reset motherboard BIOS settings to default, maybe some RAM settings messed up with GPU load.
 
Last edited:
Back
Top Bottom