Is my RAM failing?

[robot zombie]

Over this past month or so I've been having the occasional insta-restart, and much more rarely a total BSOD where everything completely locks up and I have to power off via the PSU itself.

For a while, I just didn't get it, I've had this build going for a good 5 months and past the first month of setup/tweaking, I've changed very little. I know how I like things to be set up and what generally tends to be stable. I find that point and leave it there. Too many bad experiences messing around too much and costing myself a ton of time for nothing.

The only thing new in it is a GTX 1050, which for a while I thought was the problem, as it seemed like this started after I installed it. I tried rolling back drivers. I was also running a slight OC on it, so I took that off. No change. I felt like it HAD to be the card - it always seems to crash when gaming or on youtube, so I switched to the integrated GPU on this build's A8-9600.

Still happening, though seemingly a little less frequently. I've since switched back to the GTX 1050 with no change. At some point since then, it's started crashing while I'm at work, as I come home to the lock screen every so often. And it's not wupdate doing it, every time I check there have been no new ones. Nothing else I have going auto-updates or ever necessitates an automated reboot.

Alright. Next attack point. I've had this APU pushed up to 3.9 @ 3.1ish volts for months and thought to myself that it's kind of dumb to really expect this chip to keep that up forever. Understand I don't care if I run that into the ground, as it's just a buy-in to AM4. It has always been absolutely rock-solid at this level. I was actually pretty impressed. I really put it through the ringer before settling on 3.9. But I thought okay, over time it's just buckling under that stupid-high OC - maybe these APU's just really can't handle overvolting at all, so I dropped it to 3.4 @ 1.27v. Still stable under heavy stress. Only difference is I can turn down my fan curves a little and still have things run cooler.

Between all of this I've been messing with every setting in both BIOS and Windows, just trying to find the link. Last night, my mobo finally gave me a breakthrough. I BSODed hard in the middle of a youtube video... ...another one where I had to cut power. When I power back up there's no POST. And on my Asus Strix B350-F, the red DRAM error light kicks on. First thing I do is pull out the stick closest to the CPU. One thing noteworthy here, is that it feels much warmer to the touch than the other 4GB stick installed. Not hot, just warm enough to have me wondering. I know they generate heat, but one's generating noticeably more. Anyway... this works. It POSTs and boots as per usual. I go ahead and try reinserting the stick in the same DIMM, just to confirm that it's the source of the problem.

But it boots! This has me thinking it may be the mobo. I went ahead and switched the active DIMM slot pair to see if maybe it's that slot failing. So far, so good, but we'll soon see how good. I'll be sure to report back.

But in the meantime... which is it? Failing RAM stick or mobo problems? Any way to test that or is it just wait and see? Mostly just wondering if others who have actually experienced RAM failure can chime in here. I've never had one fail on me, so I don't know what the failure progression looks like. I'm debating on whether to grab another stick of this HyperX 2400 RAM just in case, so I'm not suddenly stuck running on 4gb of RAM while I wait. Stuff aint cheap for what it is and I can't do much with one stick if neither of the current ones don't fail. Just means I'll have to buy another and upgrade to 16gb. Which is great and all, but a lot to spend for, in my case, little to no performance gains.

The only other cause I can think of... ...and it's related, is my timings. I have all of my primaries pushed down a tick. No speed OC, no voltage kick. I know from fiddling with this before that going too low causes very similar stability issues in this build, but the thing is... ...like my CPU OC, I've been running the timings this way since long before this started. Doubt it's the issue but figured I'd mention it.

 

[eidairaman1]

Wall of text, so you can see the BSOD then? Does it have the 0x code?

 

[John Naylor]

Stop Error Code ... 0x 000000124 or something like that ?
Event Viewer Errors or Warnings ?

 

[robot zombie]

Wall of text, so you can see the BSOD then? Does it have the 0x code?

Yeah, sorry bout that. Last time it happened it was so late I forgot to take it down. If/when it happens again I will be sure to. I'm trying to remember. I was pretty much asleep when it happened.

Stop Error Code ... 0x 000000124 or something like that ?
Event Viewer Errors or Warnings ?

But no, nothing like that. Kernel something...

Dammit I wish it would happen again now.

 

[John Naylor]

My son's team plays a lot of beta games and it's not at all uncommon to get BSODs from some games. But if ahppening feqyently, it likely points to another issue. Event Viewer logs are still there unless you manually cleared them. An d yes, there may be an explanation with words but the 0x0000 number is the important part.

 

[robot zombie]

I went into event manager and pulled a few more recent ones up. Don't ask me why I hadn't thought to do that.

Last nights was 0x00000000. The one before that was 0x0000007f. Going back a little further I frequently see 0x00000124 and 0x00000139. The former seem to point to what I suspected - hardware. The latter probably had something to do with my overclocks... ...that or I do have some driver issues after all. That's about all I got. Having a rough time actually referencing these codes in any meaningful way. It's either programmer level stuff or uselessly rudimentary tech support crap like uninstall this, update drivers, etc.

 

[eidairaman1]

I went into event manager and pulled a few more recent ones up. Don't ask me why I hadn't thought to do that.

Last nights was 0x00000000. The one before that was 0x0000007f. Going back a little further I frequently see 0x00000124 and 0x00000139. The former seem to point to what I suspected - hardware. The latter probably had something to do with my overclocks... ...that or I do have some driver issues after all. That's about all I got. Having a rough time actually referencing these codes in any meaningful way. It's either programmer level stuff or uselessly rudimentary tech support crap like uninstall this, update drivers, etc.

Let me look, are you on W7, 8, 10?

7F= Memory
http://aumha.org/a/stop.php#0x7f

124 is whea failure-hardware
https://docs.microsoft.com/en-us/wi...er/bug-check-0x124---whea-uncorrectable-error

0x139 could be a driver or malicious code

https://docs.microsoft.com/en-us/wi...bug-check-0x139-kernel-security-check-failure.

Lets start first by doing a clear cmos, then set factory defaults.

Then utilize an updated antivirus scanner followed by a Scan from Malware Bytes Antimalware and Super Antispyware.

Do a SFC/SCANNOW aswell.
If the bsods stop suddenly and you do not find anything malicious, it was most likely unstable overclocking of the APU and memory. If your scan finds something-you had malicious code-combination of unstable oc and malware/virus.

Get memtest86 too

Are you sure it is at 3.1Volts for vcore?

Also welcome to the world of electronics-everything can fail suddenly or show signs of impending doom. Make sure you have the ram in the correct slot for configuration with 2 sticks only.

 

[robot zombie]

Let me look, are you on W7, 8, 10?

7F= Memory
http://aumha.org/a/stop.php#0x7f

124 is whea failure-hardware
https://docs.microsoft.com/en-us/wi...er/bug-check-0x124---whea-uncorrectable-error

0x139 could be a driver or malicious code

https://docs.microsoft.com/en-us/wi...bug-check-0x139-kernel-security-check-failure.

Lets start first by doing a clear cmos, then set factory defaults.

Then utilize an updated antivirus scanner followed by a Scan from Malware Bytes Antimalware and Super Antispyware.

Do a SFC/SCANNOW aswell.
If the bsods stop suddenly and you do not find anything malicious, it was most likely unstable overclocking of the APU and memory. If your scan finds something-you had malicious code-combination of unstable oc and malware/virus.

Get memtest86 too

Are you sure it is at 3.1Volts for vcore?

Very helpful. Thank you. I will start there when I have some time. MBAM I've actually been meaning to install anyway. I use Avast by itself atm.

And no, not 3.1 lol. I'd really have problems then, heh. In my head I always seem to exclude the "1." because that never changes. I mean 1.31 o.o

Oh, and its Windows 10.

 

[eidairaman1]

Yes please keep us updated.

In my years of experience, I had my AthlonXP rig failing left and right despite a clean os install, low and behold a DDR DIMM failed.

 

[robot zombie]

Welp, just had another auto-reboot. Just browsing the internet. No youtube even. This time the BIOS hung on the post screen when it came back up. Rebooted via the power button and it posted successfully.

Starting to think it's gotta be hardware related. Went ahead and reset to factory. The only things I can think of are the really lean APU OC or the fact that I pushed my primary timings 1 tick back at stock speed. We shall see.

Nothing of note in the logs on this one.

 

[eidairaman1]

Welp, just had another auto-reboot. This time the BIOS hung on the post screen when it came back up. Rebooted via the power button and it posted successfully.

Starting to think it's gotta be hardware related. Went ahead and reset to factory. The only things I can think of are the really lean APU OC or the fact that I pushed my primary timings 1 tick back at stock speed. We shall see.

Nothing of note in the logs on this one.

Just continue to monitor,

Btw what brand and model psu aryou using?

 

[robot zombie]

Just continue to monitor,

Btw what brand and model psu aryou using?

EVGA Supernova G3 550.

 

[robot zombie]

The other day, I ran sfc and it did pick-up/fix some issues. Also setup MBAM, which picked up a couple of minor infections, all traces. Nvidia put out a new driver for the 1050, so I grabbed that. And of course this is all on stock BIOS settings - no overclocks whatsoever.

The auto-reboots seem to have stopped, or at least, I'm past due now. I'm thinking it was probably the OC causing a lot of that. Probably the more common BSODs too. Either I changed a setting at some point and forgot, or there's some degradation somewhere. It was stable for months. I wonder if the addition of the video card might've affected the stability of my CPU overclock. Just a thought. Still... I'm not sure they're meant to be overclocked to 25% over stock clock. Not a lot of info on it. Before I settled on the OC I've run these past few months, I ran it successfully at 4.1ghz with like 1.41v going to it... ...for a couple of weeks. Just because I didn't think it would do it, really. Wonder if that might've tipped the first domino. I know it's an older architecture. And even Ryzen isn't supposed to be run at 1.4v.

I'm getting a Ryzen soon. I'll be much nicer to that.

I'm rambling again. This is all well and good, but I did get another BSOD tonight. 0x0000007f :/

I have yet to run memtest. I'll be curious to see what happens. Depending on how that goes, I also have some DDR4 2133 in an older basic build in the living room. Might swap out.

 

[eidairaman1]

Ok 7f could be ram or if ocd the imc could be tripping.

 

[robot zombie]

Ok 7f could be ram or if ocd the imc could be tripping.

I'm not running any OC, so I gotta figure it's the ram. Though something has me thinking it might be more motherboard related. I have fewer problems than before, but some things are dying hard.

A thought crossed my mind. I flashed the bios when I got my video card. I can't recall exactly when this all started but it may plausibly have been right after that. Maybe there are latent compatibility issues between the new firmware and my ram? Usually the opposite happens with bios updates but who knows?

Anywho, the rabbit hole deepens - another anomaly tonight.

I just finished running a scan in MBAM. It asked me to reboot, as it has since the beginning to remove these two odd recurring traces that for whatever reason it keeps failing to remove (this one I will look into myself as I believe it's not related to my current issue... ...for a while I was running Emsisoft instead of MBAM and Avast and I'm wondering if they're post-uninstall leftovers from Emsisoft's quarantine.) Definitely strange and in need of investigating. Maybe it's something more sinister but for now I'm assuming not. I never even noticed until I looked at the logs.

The important thing is that after that reboot I completely failed to post. The dram light was cycling between orange and red. I could then reset by holding the power button to get the bios to load in safe mode. Everything looked good (completely vanilla, in fact - there's no reason for me to change anything.) But after quitting the bios I go back to that same loop of trying to post and failing, with the light cycling through each attempt. This is without me doing anything. It will just keep trying, and failing to post over and over again. I went through this watching it cycle and manually rebooting into safe mode 3 times. Just completely hung up on something.

So, I figured clearing the cmos and starting fresh was the only way. But before I would do that I had a dumb idea. I completely powered off and swapped dimm slots from A2 and B2 back over to A1 and B1. And for reasons I'm not fully grasping, it worked... ...booted right up, no need to run setup, even though the previous boot failed. I'm feeling kind of dumbfounded by it. Usually just swapping the ram slots would force me into the bios to configure what it thinks is new hardware. And if not that, it still should have forced me to configure because of the previous boot failure. <_<

The only thing I think I can say for sure at this point is that it is hardware related. And it seems it may not be as simple as bad ram, though you did suggest that bad ram can cause strange problems. You also mentioned the PSU, which although I do have a decent one, maybe still can't be ruled out.

Tomorrow I will finally run memtest and likely swap the ram in this machine for the 8gb of 2133 in the living room build. It'll be interesting to see if anything changes - for either machine. I thought for a while that I was getting somewhere and with your help some problems have definitely gone away, but right now I just don't know what to think.

 

[eidairaman1]

Since you now tell us you have a spare system (should of said earlier) you could of swapped ram then, get the known good ram and try it in the problematic system, set factory defaults in cmos, restart, then set optimized defaults.

 

[robot zombie]

Hey now, I mentioned it yesterday!

No, but you're right, it's the sensible thing to do. But had I mentioned it sooner, I wouldn't have wound up solving some other problems I had going that actually turned out not to be related like I thought they were. Some of the stuff you suggested not knowing I could swap was actually helpful to me.

For now I've decided to skip testing and just do the damned swap lol

 

[eidairaman1]

Ok good, just make sure to do what i said earlier after swapping ram. If the problems stop totally then problem solved.

Make sure to use slots a2 and b2.

By the way you may want to check out the board makers qvl for memory, another point

Have you updated the mobo bios?

 

[Hood]

I had the same problem on an i5-2500K system I built for my cousin (random shut-downs, often while idle). I discovered it was a memory problem when it shut down on me, and Windows indicated a possible memory issue. It had 2 x 4GB Corsair 1600MHz sticks. running MEMTEST86 showed many errors on 1 DIMM, so I pulled it and sent it back for RMA. Problem solved.

 

[Assimilator]

Lowering your RAM timings is an overclock.

 

[robot zombie]

Ok good, just make sure to do what i said earlier after swapping ram. If the problems stop totally then problem solved.

Make sure to use slots a2 and b2.

By the way you may want to check out the board makers qvl for memory, another point

Yeap, all done. And yes, I can confirm that these exact 4GB HyperX Fury DDR4 2400 modules are indeed supported according to the qvl. I made sure of that before I bought them, but I also did double-check the other day. I never really suspected compatibility all that much. This board has a rather long list of compatible modules... ...and that aside, this memory ran fine for almost half a year. In fact, this is the first time I've experienced ANY real stability issues with this build.

Have you updated the mobo bios?

I make it a point to do that every so often. Always have after adding anything. Last update was whenever I got my GTX 1050, which had to be just over a month ago. Right now it is current. If this swap doesn't work out, I suppose rolling back will be my next option. Looking at the timeframe for when this all surfaced, it could very well be a bum bios update.

I had the same problem on an i5-2500K system I built for my cousin (random shut-downs, often while idle). I discovered it was a memory problem when it shut down on me, and Windows indicated a possible memory issue. It had 2 x 4GB Corsair 1600MHz sticks. running MEMTEST86 showed many errors on 1 DIMM, so I pulled it and sent it back for RMA. Problem solved.

Ugh... ...don't tell me that. I'm several months outside of newegg's return window. I'd have to try my luck with Kingston.

But more and more it's seeming like I'm probably right. Somewhere around here I have a USB stick loaded up with memtest, but I'm feeling hesitant to find it at this point. :/

Lowering your RAM timings is an overclock.

Indeed it is. I don't view it in the same way as CPU or GPU overclocking, so I distinguish them. But if you consider overclocking as running something beyond spec, then yeah, it counts. And I know it can easily cause stability problems... ...more so than others, in some ways.

But that's beside the point that I am running zero overclocks right now and am still having similar problems with one particular BSOD and the occasional failure to post. I only wish it was just an OC becoming unstable.

 

[Hood]

Ugh... ...don't tell me that. I'm several months outside of newegg's return window. I'd have to try my luck with Kingston.

Most RAM has a lifetime warranty, so Kingston should cover it, their RMA page looks similar to Corsair's. Corsair wanted me to send both sticks back, but I talked them into letting me RMA only the bad one, so my cousin wouldn't have weeks of downtime. My cousin actually told me not to bother replacing it, because he couldn't tell the difference anyway with only 4 GB RAM (I explained that one day he'd need it). I told him it's no trouble, $5 postage and wait a few weeks. Maybe Kingston will do the same for you.

 

[robot zombie]

Most RAM has a lifetime warranty, so Kingston should cover it, their RMA page looks similar to Corsair's. Corsair wanted me to send both sticks back, but I talked them into letting me RMA only the bad one, so my cousin wouldn't have weeks of downtime. My cousin actually told me not to bother replacing it, because he couldn't tell the difference anyway with only 4 GB RAM (I explained that one day he'd need it). I told him it's no trouble, $5 postage and wait a few weeks. Maybe Kingston will do the same for you.

Well that's good to hear. It'd be pretty sweet if I could convince them to do that. I'll definitely be going that way if I can prove that one of these sticks is actually faulty.

Funny story. In my living room build, I ran it for like a week with only 4gb of ram and was none the wiser. I only just happened to catch it showing only 4gb in the task manager... ...turns out it was just barely not fully seated. Like, that last 16th of an inch loose. For just doing basic tasks and even some multitasking in W10, I guess 8gb isn't really needed

Only kinda joking. I was pretty surprised that I didn't really notice, though that really is just a bare-minimum, basic needs build. It exists only to play music/movies and serve as a guest computer when I have people hanging out there. No gaming or heavy workloads. I can see how your casual user could be perfectly happy with only 4gb. One friend was so convinced it must've cost me $1000 to have such a super-fast computer out there

 

[robot zombie]

Well... this build has been perfectly stable since I swapped the ram. I even OC'd the RAM up to 2400 to match the speed of the old sticks. Completely stable. Yesterday I set some rudimentary overclocks. 3.7MHz @ 1.27v on the cpu. 160 clock boost and 200 memory boost on the gpu. I did some stress testing, short and long. prime95, furmark, time spy, superposition, and of course memtest. Everything passes. No problems whatsoever.

Here's the interesting thing... ...the living room build now likes to reset itself while I'm away. NO BSODS yet - I've checked the event logs. There aren't even any specific errors associated with the reboots. But I mean, the problem does seem to have migrated with the hardware. So we say, "Okay, bad RAM." But I still can't really confirm it. All I have right now are some vague signs pointing to a bad module. I know that build is stable. I'm also running everything on it stock, which I always have. Something is funky with one of these modules, for sure.

The thing is... memtest runs fine. I thought maybe it's a compatibility issue, so I checked the qvl on that mobo. It's supported. So then I thought maybe its a power issue. The bios shows that it's giving the right voltage, as per the ram manufacturer's spec. But I suppose I don't know for a fact that the mobo is actually giving that power consistently. It's not the best... ...just a $60 matx board. I've gone ahead and upped the RAM voltage from 1.2 to 1.22. Thinking about upping SOC voltage a hair or two. I really just don't like messing with that one because of how sensitive it is.

I mean, I shouldn't have to, but wouldn't it be interesting if this changed things?

Perhaps the RAM is on the verge and just hasn't failed yet? I just keep feeling like one day it's not gonna post anymore and that'll be that, but it's not happening...

 

[eidairaman1]

When bad ram is in the mix it can cause all sorts of problems from bsods to resets.

Resets tell me that Windows is set to automatically restart upon crashing. To get a bsod to appear you must disable windows from auto restarting upon failure.

So the problem was fixed on your rig when you swapped ram and it moved to the other machine, it is the ram there bucko.