Question Specific Game and PC Crashes + Artifacts After GPU Upgrade: 650ti to 1660 Super Mini ITX OC

[CableWaves]

Having issues with game crashes, crashes in desktop, and artifacts since upgrading GPU from a 650Ti to a 1660 Super Mini ITX OC. I've gone through a lot of steps so far to resolve so have tried to add some clarity to it all below.

TL;DR: Went from 650Ti to 1660 Super Mini ITX OC, and with brand new Corsair TX650M I am getting green/purple/blue pixels on the screen and an old game needing on 512MB GPU constantly crashes. Multiple attempts to resolve issues and nothing is working.

New GPU Installation Steps

1. Going from NVIDIA to NVIDIA, I didn't use DDU. I know there is debate about this.
2. Unplugged PC and let it sit for 20 mins. Pressed power button to ensure no residual power in system.
3. Unplugged 650Ti.
4. Plugged in 1660 Super Mini ITX OC.
5. PSU needed a 6-pin to 8-pin adapter so used an adapter.
6. Powered on PC.
7. Opened NVIDIA control panel and there was an update for drivers so updated.

System

• Dell XPS8700 (bought 2013)
• Windows 10 (64-bit OS) - fully updated.
• Intel Core i7-4770 CPU @3.40GHz
• 12GB RAM DDR3
• Dell 0KWVT8 Motherboard (BIOS fully up-to-date)
• 2 TB HDD
• 1 TB SSD
• 32GB mSATA SSD
• Initial PSU: Dell D460 AM-02 PSU - 460W
• Newly installed PSU: Corsair TX650M - 650W
• Dual monitors. Both 1920x1080, 60Hz.

The Problem(s) when using Initial PSU (Dell D460 AM-02 PSU - 460W)

• PC booted up no problem. No artifacts. All seemed 100%.
• Decided to play an old game I play online - Team Fortress Classic (TFC). I have played this game on this PC for seven years (many many hours) with the 650Ti installed with zero issues.
• TFC was released in 1999. Recommended PC requirements: 800 mhz processor, 128mb ram, 32mb+ video card, Windows XP.
• Playing ~2 minutes and TFC freezes. What was on the screen stays on the screen for approx. 5 seconds. Then screen goes black (not blue) and stays black for approx. 5 seconds. PC reboots itself.
• Sometimes when it reboots the login screen gets purple pixels (e.g. image of login screen) but not always.
• I tried GPU intensive games (DOOM 2016 and Wolfenstein II) and no issues at absolute MAXED graphics settings. So I can play modern games at full graphics with no issues.
• I had two crashes when browsing the internet also (never had crashes like these prior to installing the new GPU). When these occurred I was browsing internet, dragged a browser window from one screen to the other, then the screens froze for at max 3 seconds, then the screens went black (not blue), then the PC rebooted after about 5 seconds.
• Got some advice that my PSU was likely under powered. Bought and installed a Corsair TX650M but the problems persist as below.

The Problem(s) when using New PSU (Corsair TX650M - 650W)

• Installed new PSU and played TFC (old game). Played for ~20 minutes and crashed as detailed above (previously would crash after ~2 minutes, so some improvement).
• Decided that perhaps because I didn't do a fresh driver install that was the issue. So used DDU and re-installed fresh NVIDIA drivers.
• TFC game continued to crash and in fact went 'backwards' after fresh install of drivers, in that it went back to crashing after ~2 minutes of game play.
• Additionally, since installing the fresh drivers I now get green/purple/sometimes blue pixels on the screen from just a standard boot-up. See this image for an example.
• There is also weird behaviour with these pixels. I have dual monitors. So say A display port on GPU goes to A DVI on monitor (A to A) and B display port on GPU goes to B DVI on monitor (B to B). The green pixels were appearing on A monitor but when dragging a window to B monitor the pixels were gone, drag back to A and pixels are there again. So they were only on monitor A. I switched the cables around at the monitor end only - so A display port now going to B monitor (A to B), and B display port goes to A monitor (B to A). So if it is an issue with the A display port (or the cables) it should now be showing on B monitor, but it doesn't. The pixels continue to show on A monitor. Which suggests the monitor is faulty. But this was not (and NEVER) happening before using DDU and reinstalling the drivers and also never happened using the 650Ti. When I put the 650Ti back in there are never any of these pixels.
• I have also noticed my CPU fan now randomly spikes. I have sat watching the CPU temps on CoreTemp and the spikes do not correspond with temp increases or increases in CPU usage. The fan spikes when there is no temp or CPU usage increase/decrease. Seems completely random.
• Also after DDU and fresh install, when attempting to play Wolfenstein II, I get an error message saying my GPU has run out of VRAM and I need to now reduce my settings. I messed with NVIDIA CP settings and found that when specifying the GPU as the 1660 for Wolfenstein II under "OpenGL rendering GPU" instead of leaving it set at "Auto Select" it became playable again on may settings. I only have the 1660 installed - no other GPU - and the CPU integrated graphics are not accessible when the PCI GPU is installed (at least that is my reading and there is no option to use the integrated graphics in any windows setting/device manager etc.).
• Since new PSU I have had one crash while I had browsers open but it happened while the PC was idle as I stepped away for about an hour. The windows event log said it restarted to perform a "bugcheck".

Attempts to Resolve Problems

• Verified TFC game files on Steam.
• Tried TFC compatibility setting for ALL previous version of Windows.
• Tried running TFC as administrator.
• Ran sfc /scannow: Corruption found and replaced/repaired. Crashes still occur after repair.
• Ran DISM.
• Using OLD 460W PSU: Kombustor stress test: 30 minutes, 100% GPU load, no crashes. GPU maintained 60 degrees C throughout.
• Using OLD 460W PSU: Kombustor benchmark (1920x1080, MSI-01) = 1960. Which seems around the same as other benchmarks listed on the Kombustor score index website.
• Using OLD 460W PSU: CPU temps: During stress test CPU temps on CoreTemp were in the high 50 degrees C.
• Using OLD 460W PSU: CPU temps: During Kombustor benchmark the highest CPU temps were 55-57 degrees C.
• No memory errors using Memtest86 (4 passes).
• Case is clean - looks reasonably close to brand new inside. Don't believe dust is an issue.
• Ran DELL boot up diagnostics - all clear.
• Reinstalled drivers three times. All this did was create the pixel problem.
• Windows is fully up-to-date.
• BIOS is fully up-to-date.
• Purchased new PSU (went from 460W to 650W).
• Windows Event View keeps just showing the below around the time of the crashes. I have read that no dump file being created can be the result of clean-up software (e.g CCleaner) but I have no such software installed:
  • Error - volmgr - 161 - None
    • Dump file creation failed due to error during dump creation.
  • Critical - Kernel Power - 41 - (63)
    • The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.
  • Error - EventLog - 6008 - None
    • The previous system shutdown at 15:02:37 on ‎31/‎10/‎2020 was unexpected.
  • Error - EventLog - 1101 - Event Processing
    • Audit events have been dropped by the transport. 0

Wondering if there is anything else I can do to try and sort this out? Really appreciate any advice at all.

I have an old PC I can hook up to play the old game if it's likely just incompatible drivers with the new GPU - but another person who plays this game runs it without problems on a 2070 Super, so this suggests it can't be a driver/GPU problem.

I could probably live without playing this old game on this PC since I have the old PC, but it's now the green/purple/blue pixels on startup that won't go away unless I plug in / out the monitor cables after EVERY boot up that I need to sort out.

 

[CableWaves]

Can see a request to post up UserBenchmark links on other posts. Link to my PCs UserBenchmark.

Also new behaviour, PC no booting up from shutdown and instead Dell logo with white swirling circles just going round and round. Booted up OK after holding down power button and restarting.

 

[dotas1]

Wow, you have been thorough with your post and with the steps you took.

I would suggest to try the new GPU in another system to see if any problem occurs. Artifacts are signs of a GPU going bad. That's why I suspect you got a bad card.

Very good decision to swap the PSU for 2 reasons.
First is that prebuilts almost always have crappy PSUs not meant to support anything than the components already in the system and since yours is from 2013, degradation had already occured.
Second is that you needed to use a 6pin to 8pin adapter. Adapters are almost always a no go. If a PSU does not have the necessary cables then it most probably can not support the power draw required, which can lead to damaging and/or fire (among others).

 

[CableWaves]

Thanks for suggestion dotas1. That's unfortunately the only check I can't do as everyone I know has laptops. The old PC I have is 15 years old, is a mini-tower, and the PSU wouldn't even come close.

The only final thing that's crossed my mind is the motherboard chipset. I can't locate an entry in Device Manager with the word "Chipset" in it. I can only see what I think might be the chipset listed as "Intel(R) 8 Series/C220 Series SMBus Controller - 8C22".

I had downloaded the Dell driver utility and it scanned and said everything is up to date. However the most up-to-date chipset for my motherboard is 10.1.1.7 and within that SMBus Controller the driver version is 10.1.1.38.

I've also tried the Intel® Driver & Support Assistant and this also states no updates required. However, although it lists the BIOS, it doesn't list the chipset at all, so am wondering what might be going on.

I might pop a post in the motherboard section to just ask that specific question about how to identify the chipset for my motherboard. Conscious of cross posting so will just keep it to that specific question.

 

[dotas1]

I don't think chipset is the issue here, since the BIOS is almost always to blame in such cases and you did update it. Sure, you can ask for more info in another thread about motherboard chipset as I do not know much about it.

There is another thing that you can do to see if the GPU is problematic. If you underclock memory clock of GPU and the artifacts go away then it's a very clear sign that GPU is to blame. Sometimes you might need to underclock it just a little, sometimes more than a little and in rare cases you also need to underclock core clock as well. It's very easy to do that with MSI afterburner and there are plenty of video guides on how to do it.

If you want to be sure, you can go to a repair shop and ask them if it's possible to test your GPU in one of their systems. If they see artifacts too, then RMA it asap.

 

[CableWaves]

Cool, I'll give underclocking a go later today and see what happens!

We're in lock down for a few weeks at the minute thanks to covid so unfortunately won't get to go near a PC shop. Bummer.

 

[CableWaves]

Tried underclocking Memory Clock on its own, then Core Clock on its own, then both together.

No change. Pixels are still there.

I tried small increments and then large increments (-502 was the furthest I could go).

Just in case I did it wrong, while I hit "apply" on MSI afterburner, I didn't restart the PC or anything. Should I be underclocking and then restarting for the changes to take effect?

If there was no change does this suggest the GPU isn't at fault?

I'm perplexed at this stage.

I've put the old 650Ti back in and all is well with the PC - no pixels.

c'est la vie

 

[dotas1]

Tried underclocking Memory Clock on its own, then Core Clock on its own, then both together.

No change. Pixels are still there.

I tried small increments and then large increments (-502 was the furthest I could go).

Just in case I did it wrong, while I hit "apply" on MSI afterburner, I didn't restart the PC or anything. Should I be underclocking and then restarting for the changes to take effect?

If there was no change does this suggest the GPU isn't at fault?

I'm perplexed at this stage.

I've put the old 650Ti back in and all is well with the PC - no pixels.

c'est la vie

No, you just hit apply as you did.

Let's get to basics again. Please do the following to see if anything changes:

• Reseat the GPU in the PCI slot.
• Make sure ALL cables are firmly inserted in the correct slots.
• Try botting with only the OS drive connected and not any extra drives.
• Try using only 1 RAM slot, let's say 2nd one, then just with the 4th, then both.

As for the CPU fan spiking, are you sure you did not have that before? I am asking because now that you are carefully watching how your system reacts, you will notice it for sure, but not if you weren't paying so close attention.

For Wolfenstein running out of VRAM is a known "issue" from my google search.

Wolfenstein II: The New Colossus :: Steam Community

steamcommunity.com

Wolfenstein II: The New Colossus :: Steam Community

steamcommunity.com

If the problem only occurs in TFC then I can't help but wonder if the game has issues with newer GPUs and crashes the system (very weird though).

 

[CableWaves]

Thanks again for the suggestions. I'll go through those basics again - I'm a bit tied up work wise for a bit so might not get to that for a few days but will definitely report back when I do.

As for the CPU fan spiking, are you sure you did not have that before? I am asking because now that you are carefully watching how your system reacts, you will notice it for sure, but not if you weren't paying so close attention.

I was thinking the same, that I'm now super attentive to any noise at all so am now picking this up whereas before I was just used to it. But I've been using this PC for work for 2 straight hours since putting the 650Ti back in and the fan hasn't spiked once. It would have spiked every 20 or so minutes with the 1660 plugged in. It was definitely the CPU fan spiking and not the back panel fan or GPU fan.

The spikes would occur just at idle - though I know loads of stuff can be going on in the background. I sat with CoreTemp + HWiNFO64 + Task Manager all open to see what might be going on process wise and temperature wise and mhz wise. The spikes would happen without any increase or decrease in processes, temps, mhz usage. Seemed completely random.

I do have a micro ATX motherboard, so the GPU is approx 3cm below the CPU fan/radiator. My recall is failing me but I think at idle the 1660 would sit around the high 50 degrees C and the GPU fan would not come on until it hit 60 degrees which only ever happened during game play.

I wonder if maybe heat from the 1660 at idle was affecting the CPU temperature being so close to it? Any temperature impacts from the GPU on the CPU though weren't being registered on CoreTemp (maybe they happen too quickly to register and the fan spike quickly brought the temp back down?) This could explain why the 650Ti doesn't impact the CPU fan as at idle the 650Ti sits at 32 degrees C?

For Wolfenstein running out of VRAM is a known "issue" from my google search.

I got this one sorted by specifying the 1660 as the OpenGL GPU within NVIDIA Control Panel. Reinstalling the drivers seemed to set this to "Auto Select". There is only one GPU so not sure what the control panel was 'selecting' when it was set to auto.

If the problem only occurs in TFC then I can't help but wonder if the game has issues with newer GPUs and crashes the system (very weird though).

Yeah it's definitely weird as I posted on a TFC forum and another person has a 2070 Super and has no issues playing the game.

If I wasn't getting the artifact issue I'd live with it as just some off bug with this game. But since I can't seem to rectify the artifact issue that's concerning me the most.

The odd thing here is that initially I only got artifacts after TFC crashed. But then I used DDU to clear out all GPU drivers and then re-installed them to see if that stopped the TFC crashes but since doing this fresh install I now get artifacts when just booting up (though not always - roughly half the time). When I get the artifacts on bootup now I also notice that the startup takes much longer - approximately 20 seconds longer. I know that's not a long time but it's long compared to what I'm used to and the longer boot-up coincides with the artifacts.

Thanks for your time on this and for Googling some of the issues. I'll definitely come back once I've re-tried the basics again.

Probably doesn't add much but here's the UserBenchmark with the 650Ti.

 

[CableWaves]

Let's get to basics again. Please do the following to see if anything changes:

• Reseat the GPU in the PCI slot.
• Make sure ALL cables are firmly inserted in the correct slots.
• Try botting with only the OS drive connected and not any extra drives.
• Try using only 1 RAM slot, let's say 2nd one, then just with the 4th, then both.

Will line this up to do this weekend. Just before doing it I just wanted to check that I'll have the process down correctly in terms of the RAM checking? So basically boot up and play the game / check for artifacts with:

1. RAM in slot #2 only
2. RAM in slot #4 only
3. RAM in slot #1 only
4. RAM in slot #3 only
5. RAM in slots #2 and #4 only
6. RAM in slots #1 and #3 only
7. RAM in all slots

Is there any programme I can install that might do a better job recording events leading up to a crash than Windows Event Viewer? From what I've read the errors that I see in Event Viewer are only highlighting that the PC shut down unexpectedly and are not capturing what is happening with the GPU/CPU/RAM etc. leading up to the crash.

 

[dotas1]

It's a good order, just don't use the same RAM for numbers 1/2/3/4.

I don't know any programs to record that kind of events, I can't help you there, sorry.

The only other thing that comes to my mind is to try a clean windows installation, which can be done in another drive if you don't want to miss anything from your current setup. Disconnect every other drive and install to the one left.

 

[CableWaves]

I'm beyond swamped so haven't been able to try those RAM checks. I at least previously tried memtest86 that showed no errors and I was having no issues prior to the new GPU install.

I'm 2 days from 30 days of having bought the thing so am unlikely to get to do any more tests. I think it's going to have to be an RMA.

I gave DDU and a reinstall a go, and initially it worked - no pixels, and played the game without a crash. But all the issues came back after I shut the PC down and restarted.

In relation to the pixels.

• They don't appear on a screenshot - but I've found different views on what this means. Some say if not on screenshot it can't be GPU but others seem to say differently.
• The pixelation goes away once I plug out and plug in the cable at either GPU or monitor. This suggests a bad connection. However, the pixelation could be visible but when I restart the PC without touching the connections the pixelation can be gone even though I wouldn't have touched the connection at either end. Then I could restart again and the pixelation is back. The DP connections have those two little hooks to secure them in place and the DVI connections are screwed/secured in as tight as they can go.
• The pixelation only appears on one monitor out of three tested - which would suggest it's the monitor. I've tried the same cable on all three monitors - pixelation only occurs on the one monitor so can't be the cable either. But the 650Ti doesn't cause any pixelation, only the 1660 does.
• I notice that every time the pixelation occurs the PC is slow to bootup - it's slow to both get to the initial Dell logo (slow to post?) and slow to then get to Windows. The PC is never slow to bootup with the 650Ti. If this wasn't happening (slow bootup) I'd say there is just some random interaction with the GPU and this one monitor but since it is always preceded by the slow bootup it seems like something is happening the PC when this GPU is plugged in.

All in all it's been a bit frustrating and I'm not sure to get another of the same GPU again or maybe drop to a 1650 in case my system for some random reason just doesn't play well with a more powerful GPU.

 

[CableWaves]

I might have sourced a solution for the game crashes.

Found some similar experiences with Dell XPS8700 machines and the 16** series of NVIDIA cards. Interestingly one post had a very similar issue with older games.

The solution appears to be power management related based on the below four steps appearing to be the solution based on both of these posts. Of note, these four steps appear to resolve the issue even for those with the original Dell PSU that's 460W.

1. Settings > System > Power & Sleep > Additional Power Settings > Set to High or Ultimate Performance.
2. In the sleep/turn off settings > set both to Never.
3. Under the advanced power settings > Ultimate performance > change plan settings > set PCIe Link State Power Management to off.
4. NVIDIA control panel set Global settings/Power Management to High Performance

I wanted to test each of these steps individually and found that turning off the PCIe Link State Power Management has resolved the issue by itself. Got through almost 2 hours of TFC last night and no crashes.

I wonder if by me now having the 650W PSU I don't need the other steps, whereas those with the 460W PSU need to make sure the ultimate power performance settings are turned on.

Could it be that the older game took so little out of the GPU that the PCIe was shut off which was causing the problem? Or was it shut off and then, for some reason, came back on that it caused instability?

Also, CPU fan spikes have now stopped.

I have had one occasion of the pixels at start up which continue to go away when the monitor is turned off and on. So it is perhaps something specific to this monitor as switching the cables continues to only show it on this one monitor.

 

[dotas1]

I must have missed your answers, I am sorry. Glad you partially fixed it, I couldn't help more than that.

As for the monitor, if you have another, or a tv, or you can borrow one from a friend/family to check if it happens there as well, that would help.

 

[CableWaves]

No worries at all - imagine it's hard to keep track of all the questions!

The pixelation doesn't occur on two other monitors, though it occurs randomly at startup so I can't tell if it just happened to not occur when trying the other monitors.

The only tiny thing that's still concerning me is that the pixelation occurs when the PC is slow to boot up. When the PC boots up normally there is no pixelation. But every time the PC boots up slowly, the pixelation appears, which makes me feel like there is still something happening... but what that is I can't seem to figure out.

The slow boot-up can occur when the PC is turned on for the first time that day or it can occur when I restart the PC - there is no pattern. Fast boot is disabled and I always turn off PC and unplug it at the end of the day.

 

[dotas1]

It's a real puzzle this one. I would try one more thing. Buying a quality new cable to try and see if the pixelations still occur. Although, the fact that the PC is slow to boot/post when pixelation is present, bugs me the most. That shows some other kind of issue; at least to me that's what it says.

I will ask someone else to take a look at your issue.