Question PC Randomly freezes after upgrade, then won't POST until everything is reseated or after waiting a long time.

[ZachMN15]

The other day I upgraded some parts to my computer and after everything was placed and I installed the new GPU drivers, it seemed fine, until randomly crashing while playing a game. The new parts were a Gigabyte GeForce RTX 4070TI Windforce, a Corsair RM850x 80+ GOLD, Samsung 2TB 970 EVO Plus NVMe M.2, and a fractal pop xl silent case. Specifically, what happens is that randomly, (not even when playing a game necessarily), the display freezes, sometimes the application crashes a second before this freeze, but it varies wildly, all input and output is frozen as well, but the display is still on and showing the last thing open as it crashed. Rarely, it will shut itself off after this freeze, I have to turn it off manually. On only one crash has there been a blue screen; in that case there was an error code: DRIVER_IRQL_NOT_LESS_OR_EQUAL (after a quick lookup, I think this means something is wrong with some driver or memory). The crash happens when the GPU is in either PCIe x16 slot, regardless of whether or not the new SSD is in the second m.2 slot, and regardless of how much of my 4x4GB ram is slotted. After such a crash, the computer usually does not POST (no diagnostic LED [other than a line of orange, which is usually always on] or sounds) and sometimes requires reseating all RAM sticks, the GPU, and removing all USB connections, sometimes requires less. Even then it may not happen, and I may need to wait a few minutes before it will POST again; high variance with time after a crash before I can get it up and running. If there is no crash, when I reboot, everything starts up fine. If I do need to reseat everything, it seems that at least video drivers have something wrong because the windows login screen is in a lower resolution than my monitor (360p I think?) before automatically changing to the full resolution several seconds later. I have flashed the BIOS and installed the latest GPU drivers twice, once as a clean install (both result in crashes). All voltages seem to be fine (within .1 or less) in the BIOS. I have not run MemTest (86 or 64), since RAM was never an issue before the upgrades yesterday and it all shows up in BIOS and Windows (it's already 1AM, but I will start memtest after leaving for work tomorrow morning). I have not tried my old GPU (EVGA GeForce GTX 970), or old PSU (EVGA 500W 80Plus [although I don't think this would be of use because of the GPU power draw]). I have not tried reinstalling Windows (but would do on the new 2TB NVMe) or checked for Windows updates. I was able to get a really poor-quality photo of HWMoniter (v1.51) and planned what it would display as it crashed (because I wouldn't be able to expand or collapse anything else).

SPECS:

• MOBO: GIGABYTE H370M D3H GSM
• CPU: Intel(R) Core(TM) i5-9600K CPU @ 3.70GHz
• RAM: 4x4GB 2400MHz DDR4 (2 pairs of 2 mismatched brands, same speeds)​
• GPU (new): Gigabyte GeForce RTX 4070TI Windforce​
  • Driver Version: 536.23​
• PSU: (Corsair RM850x 80+ GOLD)​
• Various Drives (see below screenshot)​
• OS: Windows 10 Home​
  • Build: 19045.3086​

Below is a picture (edit: and link to picture) of HWMonitor during one such crash. Any help would be greatly appreciated. Thanks!

https://cdn.discordapp.com/attachments/180823942041174016/1123473644555874334/image.png

 

[helpstar]

to have a look what the problem could be:
run userbenchmark.com and post the http link of your result, e.g. https://www.userbenchmark.com/UserRun/28977730

Reset the BIOS by jumper clrCMOS or JBAT or similar (eventually you will have to set the boot priority correctly after that)

check windows integrity
open the command prompt as administrator and type DISM /Online /Cleanup-Image /RestoreHealth
https://www.lifewire.com/how-to-open-an-elevated-command-prompt-2618088
https://answers.microsoft.com/en-us...em-files/bc609315-da1f-4775-812c-695b60477a93

clean boot

How to perform a clean boot in Windows - Microsoft Support

Describes how to use the clean boot process to troubleshoot a problem in Windows 8.1, Windows 8, Windows 7, or Windows Vista.

support.microsoft.com

check the memory by running memtest.org usb autoinstaller (bootable USB flash drive)

check the hard drive for errors with its manufacturer´s tool

use ddu uninstaller and reinstall the latest graphics driver

 

[ZachMN15]

to have a look what the problem could be:
run userbenchmark.com and post the http link of your result, e.g. https://www.userbenchmark.com/UserRun/28977730

Reset the BIOS by jumper clrCMOS or JBAT or similar (eventually you will have to set the boot priority correctly after that)

check windows integrity
open the command prompt as administrator and type DISM /Online /Cleanup-Image /RestoreHealth
https://www.lifewire.com/how-to-open-an-elevated-command-prompt-2618088
https://answers.microsoft.com/en-us...em-files/bc609315-da1f-4775-812c-695b60477a93

clean boot

How to perform a clean boot in Windows - Microsoft Support

Describes how to use the clean boot process to troubleshoot a problem in Windows 8.1, Windows 8, Windows 7, or Windows Vista.

support.microsoft.com

check the memory by running memtest.org usb autoinstaller (bootable USB flash drive)

check the hard drive for errors with its manufacturer´s tool

use ddu uninstaller and reinstall the latest graphics driver

While running PCBenchmark, I got a blue screen, which did not give an error code (the blue screen appeared to be corrupted saying "Yo We'll restart for yo" (kind of funny looking)), and at this point I decided to reinstall windows on the new 2TB NVMe drive. Shortly after installing, I got 2 blue screens: SYSTEM_SERVICE_EXCEPTION (dxgkrnl.sys) and then UNEXPECTED KERNEL MODE TRAP. After looking these up, it seems like this is memory related. Furthermore, when running memtest earlier today, the computer froze. At any rate, I ran Windows Memory Diagnostic on this fresh installation of Windows, and it actually returned an error. Looking in the event log, I get this error from the results (Friendly View):

+System
-UserData
-Results
LaunchType Manual
CompletionType Fail
MemorySize 16301
TestDuration 237
TestCount 12
NumPagesTested 4171830
NumPagesUnTested 1255
NumBadPages 2
T1NumBadPages 0
T2NumBadPages 0
T3NumBadPages 0
T4NumBadPages 0
T5NumBadPages 0
T6NumBadPages 0
T7NumBadPages 0
T8NumBadPages 0
T9NumBadPages 0
T10NumBadPages 0
T11NumBadPages 0
T12NumBadPages 2
T13NumBadPages 0
T14NumBadPages 0
T15NumBadPages 0
T16NumBadPages 0

I did this after doing the cmd DISM /Online /Cleanup-Image /RestoreHealth stuff and a few other related commands I found online. Does this mean that at least one of my RAM sticks or slots is bad? If so, what are some reasons that replacing seemingly unrelated parts like PSU and GPU would cause this to go bad now? Am I even on the right track by thinking RAM is the issue?

EDIT: after running Windows Memory Diagnostic again with different configurations of sticks in different slots, none gave errors, and after putting them all in again (in a different configuration than started however, there were no errors in that either.

 

[helpstar]

might be a hardware problem with your RAM

 

[ZachMN15]

That or the motherboard. At any rate, the problem has gotten worse. When plugging in either gpu into that motherboard they no longer display. Either "HDMI no signal" or "displayport no signal"and the gpu fans stop spinning after a bit. I tried the new gpu in an old build and I got the same issue, but not with the 970. Granted that was a 600w PSU and old motherboard, although I'm not sure if either of those would cause the problem. I'm going to a friend's house in the coming days so I can try the gpu in his build. If that works then Ill just buy a new mobo and/or ram as that seems to be the issue. Is it possible that the new 850 PSU made all these problems?

 

[ZachMN15]

Okay, after doing a lot of testing, swapping parts between old builds and a friend build. I have what I believe the issue is. The memory slots on the motherboard on bad. Specifically, only channel A is bad. More specifically, when I could get it to boot into memtest, there would be consistent bit flipping at bit 20 at no memory address in specific (e.g. expected FFFFFFFD, returned FFF7FFFD in memtest). As for the displayport issue I'm not sure what the problem is. My friend tested it on his gpu with his monitor, with my gpu in his build in his monitor and it worked fine with some hiccups. At any rate, it still doesn't work with my monitor with my computer. Interestingly, a friend had a very similar issue recently. He 'upgraded' his video by upgrading from 144Hz to a 240Hz monitor (same GPU he was using before though), and when he did that, after a while his display and computer froze, and then had POST issues. His issue was again, one of the ram slots was bad, and seemingly out of nowhere. It should also be noted that we have similar boards with the main difference being I have an H370 chipset and he has a Z370 chipset, but similar issue of putting more data through video, and then getting memory issues (his was channel B that failed though). Very strange and maybe not a coincidence. I wonder if this is a common occurrence or not. At any rate, the trend seems to be: upgrade video in some way, memory slot becomes broken.