[SOLVED] Freshly built PC and issue with BSOD

[Alessuk]

Hey guys!

My PC has the specs below:

• Asus Prime B450M Gaming/BR
• AMD Ryzen 5 3600
• 2x8Gb RAM XPG Spectrix 3000MHz
• EVGA Nvidia Geforce GTX 1650 SC Ultra
• EVGA 650W 100-N1-0650-L

Ive bought and installed Windows and after about 3 months of use started to have many BSODs (more than 2 per day), until my Windows did not turn on without having BSOD and I choose to restore the system.

For a couple of days it had no BSOD and now it started all over again. Whenever this crashes I can't read what's on the screen because Google Chrome is in the front of it... The hardware is ok, as well as the memory test and there is also no overheating.

The minidumps files are uploaded in the link below:
https://drive.google.com/drive/folders/1JRyazbwB_TPVwa-xoOgs3oR9MC5If6k6?usp=sharing

Someone could help me investigate this?

 

[white.a.drew]

Hey guys!

My PC has the specs below:

• Asus Prime B450M Gaming/BR
• AMD Ryzen 5 3600
• 2x8Gb RAM XPG Spectrix 3000MHz
• EVGA Nvidia Geforce GTX 1650 SC Ultra
• EVGA 650W 100-N1-0650-L

Ive bought and installed Windows and after about 3 months of use started to have many BSODs (more than 2 per day), until my Windows did not turn on without having BSOD and I choose to restore the system.

For a couple of days it had no BSOD and now it started all over again. Whenever this crashes I can't read what's on the screen because Google Chrome is in the front of it... The hardware is ok, as well as the memory test and there is also no overheating.

The minidumps files are uploaded in the link below:
https://drive.google.com/drive/folders/1JRyazbwB_TPVwa-xoOgs3oR9MC5If6k6?usp=sharing

Someone could help me investigate this?

In cmd as admin run this command

Sfc/scannow

 

[dorsai]

What hard drive are you using ?

Did you set memory timings and voltage manually or are you using XMP ?

 

[drea.drechsler]

....
Someone could help me investigate this?

First things simplify the question of hardware problems by just resetting CMOS (check your motherboard manual how to do that) and run the system with full-default settings. Now run SFC/scannow as suggested before. If all is fine now something's amiss with your BIOS settings.

If things continue to mess up then you either have defective (broken) hardware or a Windows install that's corrupted beyond the ability of simple command-line tools (like SFC or System File Checker) to fix. Troubleshooting a Windows install by deciphering mini-dumps is an exercise in futility when it's so easy to just do a System Reset. Just type Recovery Options in Cortans search and open the Recovery options applet. You can reset your PC with an option to either save all installed files or remove them. Or, even a completely clean install which is best and probably not much of a problem on a new system.

BTW: a brand-new Windows install that's already corrupted would suggest hardware problems and on a brand-new system that's most likely memory being overclocked outside of stability. So that's why resetting CMOS to default settings is an important first step. Once you get Windows to stop the BSOD's, then overclock memory (either with XMP or manual settings). BUT WHEN YOU DO: peform a memory stability test before going ahead with normal useage!

If you insist on getting help to decipher minidumps you'd probably be better off in a Windows10 support forum, not a hardware/CPU forum. In addition to the one on Tom's there's also https://www.windows10forums.com/ where there are several MSCE's active and online much of the time providing help and powershell scripts for a lot of common problem.

 

[Alessuk]

What hard drive are you using ?

Did you set memory timings and voltage manually or are you using XMP ?

I did not set any memory timings or voltage. I'm not using XMP also.

I'm using SSD Crucial BX500, 240GB, SATA and HD Seagate BarraCuda, 1TB, 3.5´, SATA. Windows 10 is installed/running on the SSD.

 

[Alessuk]

In cmd as admin run this command

Sfc/scannow

I've repeated this command and turns out any defective file.

 

[Alessuk]

First things simplify the question of hardware problems by just resetting CMOS (check your motherboard manual how to do that) and run the system with full-default settings. Now run SFC/scannow as suggested before. If all is fine now something's amiss with your BIOS settings.

If things continue to mess up then you either have defective (broken) hardware or a Windows install that's corrupted beyond the ability of simple command-line tools (like SFC or System File Checker) to fix. Troubleshooting a Windows install by deciphering mini-dumps is an exercise in futility when it's so easy to just do a System Reset. Just type Recovery Options in Cortans search and open the Recovery options applet. You can reset your PC with an option to either save all installed files or remove them. Or, even a completely clean install which is best and probably not much of a problem on a new system.

BTW: a brand-new Windows install that's already corrupted would suggest hardware problems and on a brand-new system that's most likely memory being overclocked outside of stability. So that's why resetting CMOS to default settings is an important first step. Once you get Windows to stop the BSOD's, then overclock memory (either with XMP or manual settings). BUT WHEN YOU DO: peform a memory stability test before going ahead with normal useage!

If you insist on getting help to decipher minidumps you'd probably be better off in a Windows10 support forum, not a hardware/CPU forum. In addition to the one on Tom's there's also https://www.windows10forums.com/ where there are several MSCE's active and online much of the time providing help and powershell scripts for a lot of common problem.

Firstly, thank you.

As I've said before, I already reset my PC/Windows and I'm still getting problems with BSOD. I already upgrade all my drivers by using the last version available on the fabricant website, also I took the PC to assistance in order to discard hardware problems (technical said that's nothing broken), but nothing is solving the problem anyway.

Almost everyday this happens and I use this PC to work from home, so its getting really annoying. I'm not a hardcore PC user, this is my first experience building and managing a PC of my own, so basically I don't know what to do. That's why I'm looking for help on forums in the internet.

I will look forward to resetting CMOS and checking the system files again, as you advise.

I'll keep this thread updated.

 

[dorsai]

I did not set any memory timings or voltage. I'm not using XMP also.

I would recommend you do so...set both manually in the BIOS and test for stability. Letting the motherboard set timings and voltage can be a recipe for disaster.

 

[drea.drechsler]

Firstly, thank you.

As I've said before, I already reset my PC/Windows and I'm still getting problems with BSOD. I already upgrade all my drivers by using the last version available on the fabricant website, also I took the PC to assistance in order to discard hardware problems (technical said that's nothing broken), but nothing is solving the problem anyway.

Almost everyday this happens and I use this PC to work from home, so its getting really annoying. I'm not a hardcore PC user, this is my first experience building and managing a PC of my own, so basically I don't know what to do. That's why I'm looking for help on forums in the internet.

I will look forward to resetting CMOS and checking the system files again, as you advise.

I'll keep this thread updated.

OK, so at this point I'm going to assume you've reset CMOS and you are running the system on full-on default settings for memory and CPU.

So let's try some diagnostics. The best way to corrupt Windows and make it start throwing BSOD's is to have some defective memory. That's why it's both critical to be operating in default conditions, and to test it first and very thoroughly. So the first is a memory stability test. So get MemTest...from here. It's free. Read the instructions, you'll start several copies, each instance using an equal portion of your memory minus about 2GB for Windows to function. Let it run AT LEAST to 200% on EACH instance. That will be two times through the test on all memory. If you have defective memory it will start throwing errors in the test and you'll be able to see it.

You can also use the one built in to Windows; just type "Windows Memory diagnostic" in the Cortana search box. It's very easy, just launch it and let it re-start your system into the test process. The only problem is it's hard to get results; you have to search through the Event log to find it.

 

[drea.drechsler]

I would recommend you do so...set both manually in the BIOS and test for stability. Letting the motherboard set timings and voltage can be a recipe for disaster.

I've got to disagree...

The motherboard uses default memory timings obtained from SPD values loaded by the DIMM manufacturer. In every case I've seen that's going to be DDR4 JEDEC spec. clocks, with timings and voltages the DIMM mfr. has certified will function correctly. When you load XMP values (also stored in SPID, but not the default JEDEC spec-compliant default values) it is getting into 'overclocking' and the territory of 'maybe it can...maybe it can't'.

I've never seen it stated the motherboard determines the timings and voltages on it's own, although there is a 'training' process that initializes memory. But that's a process handled by CPU/memory controller and is controlled by AGESA (in AMD systems) to my knowledge.

 

[dorsai]

I've got to disagree...

The motherboard uses default memory timings obtained from SPD values loaded by the DIMM manufacturer. In every case I've seen that's going to be DDR4 on-spec clocks, with timings and voltages the DIMM mfr. has certified will function correctly. When you load XMP values (also stored in SPID, but not the default DDR4 spec-compliant default values) is getting into 'overclocking' and the territory of 'maybe it can...maybe it can't'.

I've never seen it stated the motherboard determines the timings and voltages on it's own, although there is a 'training' process that initializes memory. But that's a process handled by CPU/memory controller and is controlled by AGESA (in AMD systems) to my knowledge.

The OP is complaining of stability issues so we have to assume something is not configured correctly. I've seen even an expensive Samsung B-die ram kit not set itself to the manufacturers listed timings when the BIOS is on auto. Whether that's a ram training issue or something else I couldn't say, but it was certainly not set to proper timings across the board. Without having access to the OPs ram to pull the JDEC specs off the sticks, at the minimum, I would manually set 1.35v and proceed with stability testing.

 

[drea.drechsler]

The OP is complaining of stability issues so we have to assume something is not configured correctly. I've seen even an expensive Samsung B-die ram kit not set itself to the manufacturers listed timings when the BIOS is on auto. Whether that's a ram training issue or something else I couldn't say, but it was certainly not set to proper timings across the board. Without having access to the OPs ram to pull the JDEC specs off the sticks, at the minimum, I would manually set 1.35v and proceed with stability testing.

I agree, the DIMM could be faulty. Whether because the MFR loaded bogus SPD data or bad RAM it amounts to the same thing. Instead of trying to make a faulty DIMM kit work...why not help OP make a case for getting it replaced under warranty? Or better yet, an entirely different kit that actually works as it should.

Does anybody know anything about this XPG Spectrix memory he has? I don't and I note he has a newly-built system, only about 3 months old. IMO it should perform flawlessly at defaults, no fiddling needed, no BSOD's, and perfect passes through any memory stability test you throw at it. And I do think that should be a first step before starting on any memory overclocking venture, even if using XMP.

 

[Alessuk]

3 hours after running Memtest86 the result is:

 

[drea.drechsler]

3 hours after running Memtest86 the result is:

That's good. And you so I understand...did you do a clean install, with your memory running in default (no XMP) settings and it still got BSOD's?

 

[Alessuk]

That's good. And you so I understand...did you do a clean install, with your memory running in default (no XMP) settings and it still got BSOD's?

Yes! I did a clean install, just setup boot to SSD in Bios, since I'm using and SSD and HD. I didn't change anything else.

No XMP, so I'm using the memory in 2600MHz even if supports 3000MHz. I was planning to enable XMP, but I didn't.

 

[drea.drechsler]

Yes! I did a clean install, just setup boot to SSD in Bios, since I'm using and SSD and HD. I didn't change anything else.

No XMP, so I'm using the memory in 2600MHz even if supports 3000MHz. I was planning to enable XMP, but I didn't.

NOW try this: change DIMM voltage to 1.35V, as @dorsai suggests and see if it BSOD's.

BTW, when it BSOD's what is the error it throws in the Event Log?

 

[Alessuk]

NOW try this: change DIMM voltage to 1.35V, as @dorsai suggests and see if it BSOD's.

BTW, when it BSOD's what is the error it throws in the Event Log?

Today, two BSODs:

BugCheck name: KERNEL_AUTOBOOST_INVALID_LOCK_RELEASE
BugCheck code: 0x162
Caused by: ntoskrnl.exe

That's the message you are referring for? Sorry if I mistake something, English isn't my mother language. ^^

 

[drea.drechsler]

Today, two BSODs:

BugCheck name: KERNEL_AUTOBOOST_INVALID_LOCK_RELEASE
BugCheck code: 0x162
Caused by: ntoskrnl.exe

That's the message you are referring for? Sorry if I mistake something, English isn't my mother language. ^^

OK..I can imagine you've done this but I chased down some Googles of the BugCheck name. They mostly all come back to a hardware error or a corrupted Windows. One other possible is the drive with the swap file is full but I assume you've plenty of space on the system drive since it's new. Since the problem persists through fresh installs (which tends to rule out corrupted Windows install) and you can't find any memory problems I'd have to say it could be a CPU defect that's developed.

At this point I'd suggest contacting AMD tech support and see about an RMA. Maybe others might come up with additional ideas, but I do tend to think a system should operate stable (no BSOD's) at default settings and voltages as a minimum. Raising voltage above spec at default settings isn't enough for me to say it's good hardware but you can try that if you want.

There is one other thing they've mentioned and that's to make sure you have the latest chipset drivers installed. Use only ones you get from the AMD support web site to be sure it's latest and not coming with bloat/scam/trash- ware motherboard mfr's often bundle in.

And one last thought: I'm not usually one to go straight to the PSU when strange problems present, but while it may still be new (at least I assume it was when you built the system 3 mo. ago), that EVGA 650W looks to be a pretty sketch quality PSU. Un-rated for 80-Plus efficiency (not even Bronze/white) and a super cheap US price for 650W doesn't give me a lot of confidence in it...and they don't have it either with a 2 year US warranty. I'd suggest getting hold of another known-good PSU and trying that out. In fact, you might do that before chasing down AMD tech support.

 

[Alessuk]

OK..I can imagine you've done this but I chased down some Googles of the BugCheck name. They mostly all come back to a hardware error or a corrupted Windows. One other possible is the drive with the swap file is full but I assume you've plenty of space on the system drive since it's new. Since the problem persists through fresh installs (which tends to rule out corrupted Windows install) and you can't find any memory problems I'd have to say it could be a CPU defect that's developed.

At this point I'd suggest contacting AMD tech support and see about an RMA. Maybe others might come up with additional ideas, but I do tend to think a system should operate stable (no BSOD's) at default settings and voltages as a minimum. Raising voltage above spec at default settings isn't enough for me to say it's good hardware but you can try that if you want.

There is one other thing they've mentioned and that's to make sure you have the latest chipset drivers installed. Use only ones you get from the AMD support web site to be sure it's latest and not coming with bloat/scam/trash- ware motherboard mfr's often bundle in.

And one last thought: I'm not usually one to go straight to the PSU when strange problems present, but while it may still be new (at least I assume it was when you built the system 3 mo. ago), that EVGA 650W looks to be a pretty sketch quality PSU. Un-rated for 80-Plus efficiency (not even Bronze/white) and a super cheap US price for 650W doesn't give me a lot of confidence in it...and they don't have it either with a 2 year US warranty. I'd suggest getting hold of another known-good PSU and trying that out. In fact, you might do that before chasing down AMD tech support.

Hello!

I did change DRAM Voltage on BIOS today. The default setting was 1.20V and the right voltage for XPG Spectrix is 1.35V, like you guys said.

Before changing the voltage another BSOD occurred, this time the BugCheck name is PAGE_FAULT_IN_NONPAGED_AREA. I've found in the article here that this error is related to RAM issues.

I'm hoping the problem will be resolved with this configuration.

 

[drea.drechsler]

...

I'm hoping the problem will be resolved with this configuration.

The manufacturer listed 1.35V as required, but that's for operating the RAM at it's rated clock speed. When it starts in default settings (not enabling XMP) RAM normally starts up at a JEDEC standard DDR4 clock speed that's very low, usually 2100 or 2400. At those low clocks 1.2V should function correctly with good RAM. If it's not starting at that low of a clock speed in default settings then it should set a higher voltage as needed, but as @dorsai noted previously not all motherboards are good about that. So manually setting 1.35V may be the right thing to do, hopefully that's the case.

But if your RAM needs 1.35V to operate stable at 2100, I would say it's faulty and should be RMA'd if you can. Either that or garbage memory which may just continue to degrade over time, as it apparently has in the 3 mo's you've been using it to the point it started the BSOD's.

And of course...none of that explains why you got a PASS on the MemTest86 stability test you ran. I'd get the MemTest stability test I linked earlier (similar name, but different), and also run the Windows Memory Diagnostic (type it in Cortana). More tests are better.

 

[Alessuk]

The manufacturer listed 1.35V as required, but that's for operating the RAM at it's rated clock speed. When it starts in default settings (not enabling XMP) RAM normally starts up at a JEDEC standard DDR4 clock speed that's very low, usually 2100 or 2400. At those low clocks 1.2V should function correctly with good RAM. If it's not starting at that low of a clock speed in default settings then it should set a higher voltage as needed, but as @dorsai noted previously not all motherboards are good about that. So manually setting 1.35V may be the right thing to do, hopefully that's the case.

But if your RAM needs 1.35V to operate stable at 2100, I would say it's faulty and should be RMA'd if you can. Either that or garbage memory which may just continue to degrade over time, as it apparently has in the 3 mo's you've been using it to the point it started the BSOD's.

And of course...none of that explains why you got a PASS on the MemTest86 stability test you ran. I'd get the MemTest stability test I linked earlier (similar name, but different), and also run the Windows Memory Diagnostic (type it in Cortana). More tests are better.

Good!

I'm gonna run this another test asap.

The Windows Memory Diagnostic I've ran already and didn't got any errors.

Other thing I'm looking forward is to change my PSU. I think I'm gonna switch to Corsair CX550 that's matching my budget. lol

 

[drea.drechsler]

.....
The Windows Memory Diagnostic I've ran already and didn't got any errors.
.....

That just makes it even more perplexing. I would expect them to catch bad RAM, they certainly take long enough to run through their tests.

The MemTest I linked has been very good for me as I've overclocked my RAM (I'm running a 3200 kit of GSkill at 3600 with a 3700X). I had to keep raising voltage till it would pass all tests to 200%, but now that I've got it there it's proven perfectly stable for over 1 1/2 years now.

As far as a PSU goes: they're peculiar beasts. Really the ONLY way you can feel confident is to get one that's been favorably reviewed. The best PSU you can afford is a good idea and if you think of it as an investment that will endure over several system upgrades and updates (which it should) it's easier to stomach.

 

[Alessuk]

So manually setting 1.35V may be the right thing to do, hopefully that's the case.

I think it's solved. Didn't got any BSODs so far.

I've checked in CPU-Z that my RAM is operating at 2600.

I did this MemTest you linked, the result is that both pass the test over 200% coverage.

Thanks @drea.drechsler and @dorsai !!!