Possibly busted RAM (gaggle of BSODs), looking for recommendations

[Frick]

Hey hey.

Specs in specs, but also here:

Asrock B450M-HDV
Ryzen 5 2600x
G-Skill Aegis F4-3000C16-8GISB x 2
Gainward RTX 3060ti Ghost OC
Seasonic Core GC500
Some drives

So in short, the symptoms are basically that the OS gets corrupted with varying frequence, more so lately. The Kingston NV1 drive tanked it's ass off (can't even boot into safe mode, dunno if the drive is gone or if it's just file corruption) last week, so now I'm running the old Kingston A400, and it's not really stable. Often requires a few attempts to boot (the latest messege was simply "CRITICAL PROCESS DIED"). All the BSOD's points to file corruption. Firefox is crashing a lot. So I'm thinking the RAM isn't playing nice with the motherboard/CPU. I've ran various memory diagnostics, but none of them has shown any errors. I'm not overclocked, everything is set to auto. I did try a XMP profile for a while, but that increased instability.

In short I'm getting a bit sick of it, so I'm thinking about getting some new RAM. I would also like to upgrade to a newer CPU at some point. Looking at the RAM QVL is depressing as most modules are low frequency and density, and older models that isn't avaliable for purchase anymore or are crazy expensive. Is the days of just tossing in whatever RAM you have laying around and it just works truly over?

Tips or ideas?

 

[Ware]

If it "Often requires a few attempts to boot", then I wouldn't worry about Firefox crashing a lot.
I wouldn't rush out to buy new RAM either.

"the OS gets corrupted with varying frequence, more so lately."
IMO you describing a system I consider totally unstable. By your description, I might assume that it never has been stable.
I would test each component independently of the others.
If you don't have access to another PC to complete the diagnosis yourself, take it to somebody that does.

"Is the days of just tossing in whatever RAM you have laying around and it just works truly over?"
On current AMD systems those days are gone.

 

[68Olds]

Personally I'd start by testing the RAM for errors.

MemTest86 - Official Site of the x86 and ARM Memory Testing Tool
Memtest86+ | The Open-Source Memory Testing Tool

And testing the Kingston drive's SMART data.

SSD Manager - Monitor Drive Health and Usage - Kingston Technology

 

[The King]

Are you running the RAM at XMP?

Unstable RAM can indeed cause all those issues.

I suggest runing 1 stick with XMP off and see if you can get the system stable.

 

[Frick]

Personally I'd start by testing the RAM for errors.

MemTest86 - Official Site of the x86 and ARM Memory Testing Tool
Memtest86+ | The Open-Source Memory Testing Tool

And testing the Kingston drive's SMART data.

SSD Manager - Monitor Drive Health and Usage - Kingston Technology

Done. No errors. When I tried the TPU memory test thing just now explorer.exe crashed, and all windows went black.

If it "Often requires a few attempts to boot", then I wouldn't worry about Firefox crashing a lot.
I wouldn't rush out to buy new RAM either.

"the OS gets corrupted with varying frequence, more so lately."
IMO you describing a system I consider totally unstable. By your description, I might assume that it never has been stable.
I would test each component independently of the others.
If you don't have access to another PC to complete the diagnosis yourself, take it to somebody that does.

"Is the days of just tossing in whatever RAM you have laying around and it just works truly over?"
On current AMD systems those days are gone.

Test how? Memtest doesn't tell anything. The computer's been run with drives disconnected and so on and no problems. No change when upgrading PSU/GPU to what I have now. I have encountered similar stuff before (a long time ago though, meaning memtest doesn't show errors but replacing the RAM fixed the problems). Plus I have can't really borrow a computer for weeks. As I said, memtest doesn't show anything, nor does any other benchmark or whatever. Random BSOD's mean .... random BSODs. Heavy loads can be fine, booting might not be. Clicking something somewhere crash the computer. The Kingston NV1 had more minidumps, from memory they were mostly about file corruption. Windbg points to file corruption (or programs not handling exceptions). This happens no matter the drive configuration.

A longer explanation is that the system is stable for six months or so (with different OS drives). Then random BSOD's. Reinstallation fixes it for a bunch of months. Since I got the Kingston NV1 (in march) the problem has escalated. Now that drive is gone, and the problem is still there and stays escalated.

Are you running the RAM at XMP?

Unstable RAM can indeed cause all those issues.

I suggest runing 1 stick with XMP off and see if you can get the system stable.

XMP off.

 

[The King]

XMP off.

I see your board has only 2 DIMMS slots.

I would still try troubleshooting with just 1 stick and not 2 until you find out what is causing these problems.

I good idea is to go into the BIOS and restore deafult config with just the 1 stick in the system then try to boot into windows.

 

[HD64G]

Try re-seating the RAM sticks before anything else. Then check them 1 by 1 to determine if one of them got faulty.

 

[Frick]

So reinstalling Windows, again, due to Critical Process Died on any attempt to boot. The install lasted four weeks or so. When I started it this morning there was an error messege that came up exactly when I pressed enter to get to the Windows login prompt, so I mever actually saw what it said. Then it died. Gave up and am now running one stick. Yay. In a few months it might give some more data on the problems I guess.

Last reinstall (2.5" SSD) I swapped SATA cables btw, so those should be good.

 

[The King]

You can download this program install it and run then click the Analyze button.
Scroll to the bottom and see what caused the crash or BSOD.

Resplendence Software - WhoCrashed, automatic crash dump analyzer

WhoCrashed, automatic crash dump analyzer for Windows

www.resplendence.com

If it shows memory management error then you know its related to ram etc.

 

[Frick]

You can download this program install it and run then click the Analyze button.
Scroll to the bottom and see what caused the crash or BSOD.

Resplendence Software - WhoCrashed, automatic crash dump analyzer

WhoCrashed, automatic crash dump analyzer for Windows

www.resplendence.com

If it shows memory management error then you know its related to ram etc.
View attachment 265666

They're gone. Had to reinstall. Not even chkdsk worked. "Critical Process Died" is very unhelpful. Also, in the past they have all pointed to file corruption and bad memory, but memtest has always been clean.

I am tempted to just give up and buy a mobo, RAM and CPU but ... nah.

 

[qubit]

I had a stick go bad on me. It was really subtle for years, but caused random BSODs. It eventually became much more frequent as several times a day, so I could finally troubleshoot it. Of course, it passed memtest. I nailed it down to that one stick by a process of elimination. It's easier for you, as you've only got two sticks, so remove one and see how it goes, then do the other one and see if the fault follows one of them.

 

[Frick]

Ok so the solution was to downclock the RAM a whole bunch.

tl;dr - RAM is unstable, Windows become more unstable over time. Various BSODs, a lot of WHEA stuff. Testing the memory shows no errors. One stick worked fine. Two works if downclocked.

 

[VuurVOS]

tl;dr - RAM is unstable, Windows become more unstable over time. Various BSODs, a lot of WHEA stuff. Testing the memory shows no errors. One stick worked fine. Two works if downclocked.

So both dimms works separate installed fine? I would say there is an issue with the processor (memory controller) or the motherboard (corrosion?, bad traces?)

 

[Frick]

So both dimms works separate installed fine? I would say there is an issue with the processor (memory controller) or the motherboard (corrosion?, bad traces?)

Yeah. I'm just blaming the Ryzen memory controller.

 

[Dr. Dro]

Seems odd as that RAM shouldn't be difficult to drive at all for a 2600X.

That motherboard is really poor quality. Maybe be on the lookout for a slightly better model to show up at a nice price sometime? Sucks to be forced to lower your memory speeds, unless you're pushing a very high performance kit.

 

[VuurVOS]

Yeah. I'm just blaming the Ryzen memory controller.

Can you run ZenTimings with one and both dimms installed and share the results with us?

ZenTimings v1.36 Beta Download

ZenTimings is an easy to use, lightweight, Windows application that can read and monitor the memory timings of AMD Ryzen processors on the AM4 platfo

www.techpowerup.com