7800X3D with Gigabyte X670 AORUS ELITE AX GREEN SCREEN Crashes without ERROR messages.

[xastunts]

7800X3D with Gigabyte X670 AORUS ELITE AX GREEN SCREEN Crashes without ERROR messages.​

Hi guy's!
I have been struggling with my AM5 build since I bought it early March of 2024.
The issue I'm having is hard crashes GREEN SCREEN but without an error code before auto reboot.
I'm using a 7800x3D CPU on a Gigabyte X670 AORUS ELITE AX motherboard.

I do need help guy's. I do suspect the issue to be coming from the voltages I set from BIOS. I have tried all the bios available from Gigabyte's website and the one I do feel is the most stable is version F10.
What are your voltages for the 7800x3D?
The Vsoc I range from 1.20 - 1.25v
MEMIO 1.35 - 1.40v
VPP 1.80 (stock)
VDDP 1.025v
VDD Misc 1.12v
VDDG CCD 0.950v (stock)
VDDG IOD 0.950v (stock)
I have these crashes even on a bios reset to optimize default settings.
I do use my computer for gaming most of it. I do get these crashes usually right when loading up a game or when watching a video on twitter or youtube.
The Ram is running in pure stock with crashes and also when using XMP without touching the subtimings.
The crashes can occur right when booting up a game or after having a system uptime for several hours.
Worse case scenario I blame this on the MOBO to be the culprit and I'm willing to replace to a different brand if I can't make it stable.
more about the PC SPEC.
PSU Corsair 1200watt AX
DDR5 Team Force 6000 ram 2x 16GB (32GB total) Asus TUF 7900 XTX. I appreciate any help I can get to help me solve this issue. Thank you for your support/ Regards Dave. Below are my ZT (Zen Timings)

 

[eidairaman1]

Have you ran it stock?

 

[QyInvoLing]

maybe the memory is damaged. do you have someone that can lend you another pair of mem?

 

[xastunts]

maybe the memory is damaged. do you have someone that can lend you another pair of mem?

Nothing wrong with the memory! I've done memory error checks and they all passed the test. Thanks for helping out though.

 

[Calenhad]

You do not mention temperatures at all. Have you logged with e.g. Hardware Info to eliminate those? A poorly mounted cooler can cause weird issues sometimes.

Based on your description I would say the PSU, CPU, and motherboard are the most likely suspects. You should look into replacing them one at a time to narrow down the source of the issue. PSU and CPU are easier to replace than the motherboard. They are also easier to loan from another computer if possible.

But you might be looking at replacing the motherboard, you should have at least a one year manufacturer warranty to take advantage of.

 

[xastunts]

You do not mention temperatures at all. Have you logged with e.g. Hardware Info to eliminate those? A poorly mounted cooler can cause weird issues sometimes.

Based on your description I would say the PSU, CPU, and motherboard are the most likely suspects. You should look into replacing them one at a time to narrow down the source of the issue. PSU and CPU are easier to replace than the motherboard. They are also easier to loan from another computer if possible.

But you might be looking at replacing the motherboard, you should have at least a one year manufacturer warranty to take advantage of.

Thank you for replying! I forgot to mention that I used a dual monitor setup and that when I got this error my 2nd monitor would turn green but my main black. Now I have totally disconnected the 2nd monitor and the main is only turning black and the sound would get stuck for about 5 seconds before auto rebooting. I do have another PSU I can swap in to test indeed. I will look onto that actually. Right now I am switching the connectors of the display port from port 1 to 3 as my gpu got 3 DP ports and 1 hdmi. After the DP ports have all been tested I will test it with a HDMI cable. After I've done that the PSU replacement test is next up in que. Thanks for trying to hel.

You do not mention temperatures at all. Have you logged with e.g. Hardware Info to eliminate those? A poorly mounted cooler can cause weird issues sometimes.

Based on your description I would say the PSU, CPU, and motherboard are the most likely suspects. You should look into replacing them one at a time to narrow down the source of the issue. PSU and CPU are easier to replace than the motherboard. They are also easier to loan from another computer if possible.

But you might be looking at replacing the motherboard, you should have at least a one year manufacturer warranty to take advantage of.

OH, Yes I've been monitoring all my temps with HWinfo64 and they are all at working operations idleing temps. I run a 360 AIO water cooler for the cpu.

 

[Sarajiel]

I'm using a 7800x3D CPU on a Gigabyte X670 AORUS ELITE AX motherboard.

Which revision of the board are you using? The version number is printed on the PCB in the bottom left corner below the PCIe slots near the screw hole.
There are four different variations of that board, and the v1.3 has its own BIOS branch! You shouldn't be able to flash the wrong branch, but we are talking about Gigabyte here.

Is your CPU running at stock?

I do feel is the most stable is version F10.

That's a fairly outdated BIOS version that should actually be avoided. If you want to try older BIOS versions, don't go below F20 or FA1. F31 or FB2 should be fine if your issues are memory related.

The Vsoc I range from 1.20 - 1.25v
MEMIO 1.35 - 1.40v
VPP 1.80 (stock)
VDDP 1.025v
VDD Misc 1.12v
VDDG CCD 0.950v (stock)
VDDG IOD 0.950v (stock)

From my personal experience with Gigabyte boards, set a fixed Vsoc in the BIOS at 1.2V, or 1.25V if your CPU actually needs that much for DDR5-6000.

On the newer BIOSes (F20 - F31) you shouldn't need to set any voltages outside VDDP manually, and usually that works best at 1050mV. For more ambitious overclocks, you can set VDD to up to 1.43V manually. Avoid going over that value while testing different BIOS versions, because some don't work with high voltage mode. Both your VDDIO and VDDQ seem way too high for a Gigabyte board. They usually have a sweet spot at 1.25V for both in BIOS versions below F31.
Typically, Gigabyte's BIOS is fairly good at setting most voltages and all terminations automatically for DDR5-6000 on more current AGESAs (v1.1+).

Judging from your ZenTimings screenshot, the most likely culprit seems to be an unstable CPU overclock or your tREFI is too high for the temperature in your case, assuming that you don't have a hardware defect.

I've done memory error checks and they all passed the test.

Which ones and how many days did you run them?

 

[xastunts]

Open case while using it! Very good airflow with this Cooler master case. The 2X fans at the front are HUGE!

Which revision of the board are you using? The version number is printed on the PCB in the bottom left corner below the PCIe slots near the screw hole.
There are four different variations of that board, and the v1.3 has its own BIOS branch! You shouldn't be able to flash the wrong branch, but we are talking about Gigabyte here.

Is your CPU running at stock?


That's a fairly outdated BIOS version that should actually be avoided. If you want to try older BIOS versions, don't go below F20 or FA1. F31 or FB2 should be fine if your issues are memory related.


From my personal experience with Gigabyte boards, set a fixed Vsoc in the BIOS at 1.2V, or 1.25V if your CPU actually needs that much for DDR5-6000.

On the newer BIOSes (F20 - F31) you shouldn't need to set any voltages outside VDDP manually, and usually that works best at 1050mV. For more ambitious overclocks, you can set VDD to up to 1.43V manually. Avoid going over that value while testing different BIOS versions, because some don't work with high voltage mode. Both your VDDIO and VDDQ seem way too high for a Gigabyte board. They usually have a sweet spot at 1.25V for both in BIOS versions below F31.
Typically, Gigabyte's BIOS is fairly good at setting most voltages and all terminations automatically for DDR5-6000 on more current AGESAs (v1.1+).

Judging from your ZenTimings screenshot, the most likely culprit seems to be an unstable CPU overclock or your tREFI is too high for the temperature in your case, assuming that you don't have a hardware defect.


Which ones and how many days did you run them?

First of all thank you very much for your very detailed post! I'm running revision v1.2 on this board. And for the higher vdd and vddq voltages it is pure stock XMP for my team force ram kit. Sure I could try to run them slightly lower if that is what you might suspect could be the issue causing the crash. I've got these crashes both with PBO and -CO on and even when it was totally turned off. So for this troubleshoot I'm going to boot into bios to turn off the PBO totally for the CPU.
For the memory test I did run win11 built in memory error check that would load into bios and run the test.

Which revision of the board are you using? The version number is printed on the PCB in the bottom left corner below the PCIe slots near the screw hole.
There are four different variations of that board, and the v1.3 has its own BIOS branch! You shouldn't be able to flash the wrong branch, but we are talking about Gigabyte here.

Is your CPU running at stock?


That's a fairly outdated BIOS version that should actually be avoided. If you want to try older BIOS versions, don't go below F20 or FA1. F31 or FB2 should be fine if your issues are memory related.


From my personal experience with Gigabyte boards, set a fixed Vsoc in the BIOS at 1.2V, or 1.25V if your CPU actually needs that much for DDR5-6000.

On the newer BIOSes (F20 - F31) you shouldn't need to set any voltages outside VDDP manually, and usually that works best at 1050mV. For more ambitious overclocks, you can set VDD to up to 1.43V manually. Avoid going over that value while testing different BIOS versions, because some don't work with high voltage mode. Both your VDDIO and VDDQ seem way too high for a Gigabyte board. They usually have a sweet spot at 1.25V for both in BIOS versions below F31.
Typically, Gigabyte's BIOS is fairly good at setting most voltages and all terminations automatically for DDR5-6000 on more current AGESAs (v1.1+).

Judging from your ZenTimings screenshot, the most likely culprit seems to be an unstable CPU overclock or your tREFI is too high for the temperature in your case, assuming that you don't have a hardware defect.


Which ones and how many days did you run them?

I have 2 other memory programs Karhu and TM5. Do you recommend me to run both of them for more thorough memory error checks?

And I forgot that I had loaded a bios profile with XMP including the subtimings. This one is without subtimings all on XMP AUTO.

Oh yeah I bought this in March 2024 so my warranty hasn't expired yet. Just I need to be sure and I also need to convince Gigabyte's support that there's something wrong with the board first of all. =)

 

[Sarajiel]

And for the higher vdd and vddq voltages it is pure stock XMP for my team force ram kit.

Keep them on auto for testing. Usually, the board will use the VDD from EXPO/XMP and override the VDDQ as needed. VDDIO depends on the BIOS version, it will either be the same as VDDQ or 50mV higher, but keep that on auto as well for testing.

I'm running revision v1.2 on this board.

That means it's an 8-layer PCB, which makes this board actually one of the better boards for memory overclocking.

I've got these crashes both with PBO and -CO on and even when it was totally turned off. So for this troubleshoot I'm going to boot into bios to turn off the PBO totally for the CPU.

That's interesting in a frustrating way. My general experience with Zen3 and Zen4 tuning is that you keep the CPU at stock and tune the memory first. After intensively stress testing, you can start to tune the CPU later. Otherwise, it gets incredibly hard to pin down any errors.

For the memory test I did run win11 built in memory error check that would load into bios and run the test.

Ouch! Forget that one for overclocking.

Try to use a variety of computational and pattern stress tests for memory tuning. Good ones for AM5 are typically these:
- y-cruncher, primarily the VT3 stress test, but you can combine it with N63 and FFTv4 for longer runs, minimum 2 hours, I prefer 12+ hours
- Linpack Xtreme, a good test for infinity fabric stability and thermal cycling issues, run for multiple hours
- Prime95, run the "large FFTs" preset for multiple hours, it's a good test to see if tREFI is too long, but might not get extremely hot on single-CCD CPUs
- Karhu memory test, probably the best pattern tester, run for multiple hours
- TestMem5 v0.13.1 (from GitHub, not the old version), has multiple presets, use at least the DDR5 one for Ryzen for multiple hours
- HCI MemTest, outdated and clunky to use, paid version is almost unobtainable nowadays, the free version is fiddly to use on modern CPUs, but a decent alternative as secondary test for folks who don't want to spend money on Karhu

To check if your Vsoc is actually high enough, run FurMark (version doesn't matter) in parallel to above suggested tests for a couple of hours at least. There are more stress tests out there that you can add to the mix as well. Don't rely on a single test alone when overclocking or tuning your memory.
Personally, I try to run a combination of these kinds of tests for at least a consecutive 48 hours. It's not uncommon to see errors occur around 18 to 20 hours into these tests.

You could also try to test with a 80mm to 120mm fan pointing at your memory sticks. Just place it on the back of your GPU and see if the crashes disappear. If they do, your memory sticks are getting too hot, and you should look for a more permanent solution. Jonsbo's NF-1 memory cooler is excellent if you can get it cheap.

 

[BlaezaLite]

I've no idea about all this, but I suggest filling out your system specs for people to easily see.

 

[Caring1]

Nothing wrong with the memory! I've done memory error checks and they all passed the test. Thanks for helping out though.

Vram memory can cause green screen crashes.

 

[xastunts]

Keep them on auto for testing. Usually, the board will use the VDD from EXPO/XMP and override the VDDQ as needed. VDDIO depends on the BIOS version, it will either be the same as VDDQ or 50mV higher, but keep that on auto as well for testing.


That means it's an 8-layer PCB, which makes this board actually one of the better boards for memory overclocking.


That's interesting in a frustrating way. My general experience with Zen3 and Zen4 tuning is that you keep the CPU at stock and tune the memory first. After intensively stress testing, you can start to tune the CPU later. Otherwise, it gets incredibly hard to pin down any errors.


Ouch! Forget that one for overclocking.

Try to use a variety of computational and pattern stress tests for memory tuning. Good ones for AM5 are typically these:
- y-cruncher, primarily the VT3 stress test, but you can combine it with N63 and FFTv4 for longer runs, minimum 2 hours, I prefer 12+ hours
- Linpack Xtreme, a good test for infinity fabric stability and thermal cycling issues, run for multiple hours
- Prime95, run the "large FFTs" preset for multiple hours, it's a good test to see if tREFI is too long, but might not get extremely hot on single-CCD CPUs
- Karhu memory test, probably the best pattern tester, run for multiple hours
- TestMem5 v0.13.1 (from GitHub, not the old version), has multiple presets, use at least the DDR5 one for Ryzen for multiple hours
- HCI MemTest, outdated and clunky to use, paid version is almost unobtainable nowadays, the free version is fiddly to use on modern CPUs, but a decent alternative as secondary test for folks who don't want to spend money on Karhu

To check if your Vsoc is actually high enough, run FurMark (version doesn't matter) in parallel to above suggested tests for a couple of hours at least. There are more stress tests out there that you can add to the mix as well. Don't rely on a single test alone when overclocking or tuning your memory.
Personally, I try to run a combination of these kinds of tests for at least a consecutive 48 hours. It's not uncommon to see errors occur around 18 to 20 hours into these tests.

You could also try to test with a 80mm to 120mm fan pointing at your memory sticks. Just place it on the back of your GPU and see if the crashes disappear. If they do, your memory sticks are getting too hot, and you should look for a more permanent solution. Jonsbo's NF-1 memory cooler is excellent if you can get it cheap.

once again many many thanks for such a detailed reply! I appreciate it very very much sir!
So for the the temps I think we are more than fine as I've been monitoring them through hwinfo64 and they stays in the 50s degrees. the bottleneck won't be until they go 65.
The thing is even a reset bios and load the optimal defaults while make the pc crash by browsing videos or playing right when booting it up or when ALT TAB to another application. I have less crashes on F10 bios and with the voltages I run on my ZT.

I have no problem in running the karhu and all the other apps you have recommended me, I just feel that something must clearly be wrong here as I can't make it stable even with the bios defaults.

Vram memory can cause green screen crashes.

My GPU you say? I run the card on fast timings. Maybe I can try it on default timings mode.

 

[AsRock]

Vram memory can cause green screen crashes.

Was thinking kinda the same thing, how ever green screen has been Vulcan and black DX12. but he needs to set every thing default. Could be totally unrelated as well.

Be much better to set every thing to default see how it runs and go from there and not change multiple things at once as the error could be from anywhere.

 

[Bill_Bright]

I do have another PSU I can swap in to test indeed.

You need to do this ASAP. Since EVERYTHING inside the PC case depends on good clean, stable power, you need to ensure you are providing it.

Green screen crashes typically indicate hardware issues and crashes "without ERROR messages" are common with PSU issues. This is because a sudden interruption or failure of one or more of the supplied voltages can result in a system crash occurring so rapidly, Windows does not have time to dump and enter any error message.

So you need to swap in a known good PSU now to, if nothing else, eliminate your current PSU from the equation.

 

[TheLostSwede]

Your memory Voltages are all over the place and so are your memory timings.
This is the non RGB version of the same RAM.

 

[xastunts]

Was thinking kinda the same thing, how ever green screen has been Vulcan and black DX12. but he needs to set every thing default. Could be totally unrelated as well.

Be much better to set every thing to default see how it runs and go from there and not change multiple things at once as the error could be from anywhere.

even after a reset bios to default settings all stock default all on stock not touching anything at all. PBO off all stock voltages for the CPU and with the ram in JEDEC (NO XMP) and stock voltages would cause this crash. And The green screen only shows green when using a dual monitor setup and the green error will be shown on the secondary monitor as the main shows pitch black. Disconnecting the 2nd monitor ending up with showing only a black screen on the main monitor.

 

[eidairaman1]

Try card with different monitors in a different system

 

[Shadowized]

I do get these crashes usually right when loading up a game or when watching a video on twitter or youtube.

I'm wondering if any memory dumps have been written to the C:\Windows\Minidump folder when the system is in that frozen state, can you check? because from what I've read so far I am more inclined to think that is a GPU problem especially since you said it happens with the bios defaults. You could try using DDU and installing a different GPU driver, along with maybe downclocking the GPU core/mem to see if that would help. Aside from that testing that other PSU is also a good idea.

The green screen only shows green when using a dual monitor setup and the green error will be shown on the secondary monitor as the main shows pitch black. Disconnecting the 2nd monitor ending up with showing only a black screen on the main monitor.

These kind of issues are exactly why I was forced to moved away from AMD cards, I don't know what it is but the experience for multi-monitor users on windows with the AMD GPU drivers is just flatout terrible, I always had to turn off hardware acceleration in every program (browsers, chat apps, etc) just to prevent the card from constantly spiking the core clock/memory clocks up resulting in weird artifacts, mouse cursor corruption, screens going into a single color, and the odd system freezes, all of it was caused by the GPU drivers and as soon as I stopped using them the problems magically all went away.

 

[Sarajiel]

Your memory Voltages are all over the place and so are your memory timings.

You can't compare voltages of AGESA 1.0.0.x with AGESA 1.2.x.x directly. Especially since Gigabyte changed a lot of the automatic voltage behavior starting with the AGESA 1.2.x.x BIOSes.
As mentioned earlier, at DDR5-6000, you should not need to play with any voltages on 8-layer Gigabyte boards for any appropriate EXPO/XMP memory kits on a current BIOS. You only should touch a few specific voltages, if you want to tighten your sub-timings. Personally, I'm always surprised how many people start fiddling with voltages and terminations for timings that can be achieved on auto-settings, even on potato-grade CPUs.

@TheLostSwede, your timings are fairly horrible. Not changing tREFI ignores the setting with the single most performance increase, and your other timings are a bit all over the place as well.
Also, VDDQ and VDDIO look like you didn't enable EXPO/XMP? You should do that, because this also enables some other automatic settings for memory training, unless you want to set them all manually for some reason.

I'm wondering if any memory dumps have been written to the C:\Windows\Minidump folder when the system is in that frozen state, can you check? because from what I've read so far I am more inclined to think that is a GPU problem especially since you said it happens with the bios defaults. You could try using DDU and installing a different GPU driver, along with maybe downclocking the GPU core/mem to see if that would help. Aside from that testing that other PSU is also a good idea.

Running an integrity check in Windows might also be a good idea. If the system is crashing regularly for months, the Windows installation might have disintegrated by now.

The OP should start by running the following at least twice:

sfc /scannow

and if that keeps throwing errors, try to fix the OS with:

DISM /Online /Cleanup-image /Restorehealth

If the errors in sfc persist after that, a fresh Windows install seems to be a sensible idea.

Creating a fresh, privileged account for testing might also get rid of any corrupted data in the existing user account.

 

[TheLostSwede]

You can't compare voltages of AGESA 1.0.0.x with AGESA 1.2.x.x directly. Especially since Gigabyte changed a lot of the automatic voltage behavior starting with the AGESA 1.2.x.x BIOSes.
As mentioned earlier, at DDR5-6000, you should not need to play with any voltages on 8-layer Gigabyte boards for any appropriate EXPO/XMP memory kits on a current BIOS. You only should touch a few specific voltages, if you want to tighten your sub-timings. Personally, I'm always surprised how many people start fiddling with voltages and terminations for timings that can be achieved on auto-settings, even on potato-grade CPUs.

@TheLostSwede, your timings are fairly horrible. Not changing tREFI ignores the setting with the single most performance increase, and your other timings are a bit all over the place as well.
Also, VDDQ and VDDIO look like you didn't enable EXPO/XMP? You should do that, because this also enables some other automatic settings for memory training, unless you want to set them all manually for some reason.

Happy now?

 

[xastunts]

I'm wondering if any memory dumps have been written to the C:\Windows\Minidump folder when the system is in that frozen state, can you check? because from what I've read so far I am more inclined to think that is a GPU problem especially since you said it happens with the bios defaults. You could try using DDU and installing a different GPU driver, along with maybe downclocking the GPU core/mem to see if that would help. Aside from that testing that other PSU is also a good idea.


These kind of issues are exactly why I was forced to moved away from AMD cards, I don't know what it is but the experience for multi-monitor users on windows with the AMD GPU drivers is just flatout terrible, I always had to turn off hardware acceleration in every program (browsers, chat apps, etc) just to prevent the card from constantly spiking the core clock/memory clocks up resulting in weird artifacts, mouse cursor corruption, screens going into a single color, and the odd system freezes, all of it was caused by the GPU drivers and as soon as I stopped using them the problems magically all went away.

Many thanks for replying! Actually I have tried to set the mini dumps (windows stock) to full dump and kernel dumps all without success including the mini dumps. Perhaps that says anything to you?
I made sure im using a page file greater than my total RAM of course.

Yeah I actually ran my GPU slightly less than standards core, but on some OC for the mem. my max core is about 3000 and I ran it 2400 min and 2500 max. 2700 mem clock. I am trying just a default setting for the gpu clocks and mems now just to pin out that the problem aint coming from there.

a new psu had been swapped in right now as I post this. the old 1200X corsair got its little brother the 850 rmi corsair.

I am using the Amd adrenaling drivers directly from AMD website. not sure what drivers I should use if i clean the system with DDU.

You can't compare voltages of AGESA 1.0.0.x with AGESA 1.2.x.x directly. Especially since Gigabyte changed a lot of the automatic voltage behavior starting with the AGESA 1.2.x.x BIOSes.
As mentioned earlier, at DDR5-6000, you should not need to play with any voltages on 8-layer Gigabyte boards for any appropriate EXPO/XMP memory kits on a current BIOS. You only should touch a few specific voltages, if you want to tighten your sub-timings. Personally, I'm always surprised how many people start fiddling with voltages and terminations for timings that can be achieved on auto-settings, even on potato-grade CPUs.

@TheLostSwede, your timings are fairly horrible. Not changing tREFI ignores the setting with the single most performance increase, and your other timings are a bit all over the place as well.
Also, VDDQ and VDDIO look like you didn't enable EXPO/XMP? You should do that, because this also enables some other automatic settings for memory training, unless you want to set them all manually for some reason.


Running an integrity check in Windows might also be a good idea. If the system is crashing regularly for months, the Windows installation might have disintegrated by now.

The OP should start by running the following at least twice:

sfc /scannow

and if that keeps throwing errors, try to fix the OS with:

DISM /Online /Cleanup-image /Restorehealth

If the errors in sfc persist after that, a fresh Windows install seems to be a sensible idea.

Creating a fresh, privileged account for testing might also get rid of any corrupted data in the existing user account.

Thousands of thanks are not enough to show how much I appreciate your detailed answer, and also the others too that are trying to help me out =).
Yes I am now running the XMP with the stock voltages and not touching them manually. They have decided to go with 1.35 VDD and 1.25 for the VDDQ.
I could add that I have had issues even though right after a clean install of win 11 and also on win 10 , with only installing the win updates and installing the AMD chipset drivers with the adrenaline for the GPU and only 1 game. Battle net and call of duty warzone. Even then I do get these crashes. Maybe it is directx related or the GPU driver is just conflicting with my system somehow.
I am willing to format the drive and install a new clean win 11 again and try it all over again even though i've already been there with no success.

I'm wondering if any memory dumps have been written to the C:\Windows\Minidump folder when the system is in that frozen state, can you check? because from what I've read so far I am more inclined to think that is a GPU problem especially since you said it happens with the bios defaults. You could try using DDU and installing a different GPU driver, along with maybe downclocking the GPU core/mem to see if that would help. Aside from that testing that other PSU is also a good idea.


These kind of issues are exactly why I was forced to moved away from AMD cards, I don't know what it is but the experience for multi-monitor users on windows with the AMD GPU drivers is just flatout terrible, I always had to turn off hardware acceleration in every program (browsers, chat apps, etc) just to prevent the card from constantly spiking the core clock/memory clocks up resulting in weird artifacts, mouse cursor corruption, screens going into a single color, and the odd system freezes, all of it was caused by the GPU drivers and as soon as I stopped using them the problems magically all went away.

yeah I actually suspected that too! discord also uses hw acc and so does the browsers. i start with the PSU first to nail down that it aint a problem with it.

 

[Shadowized]

Many thanks for replying! Actually I have tried to set the mini dumps (windows stock) to full dump and kernel dumps all without success including the mini dumps. Perhaps that says anything to you?
I made sure im using a page file greater than my total RAM of course.

Just to confirm, inside that folder I mentioned, it's empty and there are no .dmp files? If there aren't any files being generated I would think it's probably on the hardware side of things, Does the Windows Event Viewer mention anything after a reboot?

a new psu had been swapped in right now as I post this. the old 1200X corsair got its little brother the 850 rmi corsair.

That 850W should be more than enough to help narrow it down.

with only installing the win updates and installing the AMD chipset drivers

what GPU and chipset driver version do you have installed currently?

 

[xastunts]

Latest BIOS FLASHED ALL STOCK BIOS

Just to confirm, inside that folder I mentioned, it's empty and there are no .dmp files? If there aren't any files being generated I would think it's probably on the hardware side of things, Does the Windows Event Viewer mention anything after a reboot?


That 850W should be more than enough to help narrow it down.


what GPU and chipset driver version do you have installed currently?

correct no dump file written.I do not know how to check windows event viewer?
I'm using amd's latest GPU and chipset drivers from their website.

I think the problem is coming from the MEM OC for the GPU. I'm running my GPU all on stock and never had a PC crash since I did so. Game crashed one time but back to windows and pc all stable after crash. So I flashed to latest bios just to see if I got rid of the crashes. ALL stock settings from bios right after the flash as you can see on my ZT on my post above this one.

 

[Shadowized]

I think the problem is coming from the MEM OC for the GPU.

That's very likely the culprit. My guess is the GPU is crashing so severely that it causes the device to disappear from the PCIe tree, resulting in your motherboard freaking out and doing a full reset to try and recover itself. Would make sense given the lack of dump files being generated, the OS has no idea because the problem is on the hardware side. I'd give it a day and then try slowly re-implementing the EXPO settings and whatever other BIOS settings to see if it remain stable.

As for the Event viewer you just open it and look under the main tree for critical errors, or navigate to Windows Logs > System and see if any events are marked as Critical/Error within the time-frame that a crash occurred.

 

[xastunts]

That's very likely the culprit. My guess is the GPU is crashing so severely that it causes the device to disappear from the PCIe tree, resulting in your motherboard freaking out and doing a full reset to try and recover itself. Would make sense given the lack of dump files being generated, the OS has no idea because the problem is on the hardware side. I'd give it a day and then try slowly re-implementing the EXPO settings and whatever other BIOS settings to see if it remain stable.

As for the Event viewer you just open it and look under the main tree for critical errors, or navigate to Windows Logs > System and see if any events are marked as Critical/Error within the time-frame that a crash occurred.

Yes I see the critical errors here.
Kernel-Power event ID 41 Task category 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.