[SOLVED] Random hard freezes (especially when idle) while running XMP profile on RAM

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
Hey guys,

I'm getting hard freezes and I've pretty much narrowed down that the XMP profile is the culprit. While doing light loads like web browsing / watching YouTube or simply after leaving the PC idle for 3-6 hours and then returning, it will hard freeze.

Mouse and keyboard will be totally unresponsive (though still maintain lights) and monitor will display the last image (including mouse cursor) but will otherwise have no activity. The clock in the bottom-right hand corner of the system tray will be frozen, etc.

Of note, whenever this happens, my RGB peripherals stop syncing with ICUE / Wallpaper engine; my front case fans stop using the given theme and go back to the default rainbow RGB cycle. That's my first giveaway if I just walk into the room and look at the PC. If it's back to using rainbow cycle loop instead of the Wallpaper Engine themed lighting, I know the PC is frozen.

This is my RAM:

Corsair Dominator Platinum RGB 32GB (2x16GB) DDR4 3200 (PC4-25600) C16 1.35V Desktop Memory
https://www.amazon.com/gp/product/B07N3LDKR1/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1
Model: CMT32GX4M2D3600C18

It seems to be supported on my motherboard's QVL list.

Motherboard is an ASUS Z490-E.

I ran 5 and a half hours of Memtest86 yesterday (all 4 passes, 13 tests each) with no errors on the RAM as it was set to the XMP profile.

Stock speed of the RAM is 2133.

I recently installed a 3080ti and a 1000w corsair PSU, but this issue pre-dates installation of this hardware. I simply tried reenabling the XMP to get the most performance out of the 3080ti and the issue remains the same.

If I leave my PC idle for hours at stock RAM settings (no XMP), then I don't get any freezes. I'm testing this once more just to be absolutely sure (I currently am at stock optimized defaults BIOS settings). If the PC somehow freezes in the next 5-8 hours, then I guess XMP isn't the actual issue, but right now I'm very confident that it is.

RAM is being detected fine in windows and I re-seated the sticks in their slots recently. They are in the appropriate dual-channel slots (A2 and B2 I believe).

I have Windows 11 installed in a new m.2 drive so I do not believe this is a SSD issue.

Windows is set to Performance mode and I have disabled all sleep / suspension features in the power management.

I currently do not have the CPU overclocked at all.

Windows 11 is fully updated. Virus scans come up with nothing. I updated my BIOS tonight to the latest version (dated 12/16/21) and this didn't help. I slept for about 5 hours (with XMP applied) and woke up to the PC frozen.

XMP I and XMP II both freeze.

Any help would be massively appreciated. All I can currently think is that my RAM just won't work with XMP applied and that I'll need to buy a new kit.

Specs:
i9 10900KF
RTX 3080 ti (water-cooled ASUS model)
Corsair Dominator Platinum RGB 32GB (2x16GB) DDR4 3200 (PC4-25600) C16 1.35V Desktop Memory
Z490-E Motherboard
Creative Soundblaster AE-7
2TB Samsung M.2 980 EVO
PSU: Corsair RMX Series (2021), RM1000x, 1000 Watt, Gold, Fully Modular Power Supply
Windows 11
 
Last edited:

Lutfij

Titan
Moderator
Out of curiosity, did you perform the ME update as well in between the BIOS updates? Might want to use Samsung's Magician app and see if your SSD is pending any firmware updates. As opposed to X.M.P, have you tried manually inputting the frequency, timings and voltage? I'd set the voltage to 1.45v to begin with and if things remain stable, dial the voltage down until you find the lock ups happening again.
 
Reactions: Stinger86

Lutfij

Titan
Moderator
Out of curiosity, did you perform the ME update as well in between the BIOS updates? Might want to use Samsung's Magician app and see if your SSD is pending any firmware updates. As opposed to X.M.P, have you tried manually inputting the frequency, timings and voltage? I'd set the voltage to 1.45v to begin with and if things remain stable, dial the voltage down until you find the lock ups happening again.
 
Reactions: Stinger86

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
ME update? I wasn't sure what this is so I googled it. Intel Management Engine update? No I haven't installed that update to my knowledge.

Good idea on the SSD firmware update. The REASON this is a new SSD is because my other m.2 gradually crapped out. It would randomly freeze but it would also sometimes fail to boot into Windows and go into a recovery loop. I knew it was toast when I tried to reinstall Windows onto it and got a red X because the hardware was done. But after I picked up the new m.2 SSD, all those issues ceased.

Everything was fine (even for a day with the new 3080 ti) until I figured my RAM speed was kneecapping my max FPS in some games and tried reenabling XMP.

I haven't tried manually inputting frequency, timings, and voltage. I'm a little unfamiliar with where & how to do this. I did try applying the XMP profiles (which of course loaded the pre-set timings and frequency) and manually bumping the voltage to 1.45 from 1.35. During this same tweak, I also tried increasing VCCSA to 1.15 and VCCIO to 1.20 per a Reddit thread I had seen where another person was having a similar issue to me. But with these settings applied, I did eventually encounter another hard freeze.

Is there anywhere I can find the XMP timings on my RAM? Do I have to physically inspect the stick itself? I'm not sure if the printed numbers would show stock or XMP timings. Where would I enter these in the BIOS?

And what difference would manually entering these values make if the values themselves are the same?

EDIT: My SSD firmware was indeed out of date. I have successfully updated it. Probably a good thing to do in any event, even if the SSD isn't the cause, so I appreciate you pointing this out. Awaiting further advisement.


 
Last edited:

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
Update:

I just got out of BIOS. I had temporarily set the XMP profile on and went to save and reset just so the BIOS would show me the values that were tweaked. In addition to voltage, frequency, and timings, the BCLK is also getting toggled from auto to 100. Not sure if that is screwing things up or not.

I reverted all the changes by loading optimized defaults. Then I went in and manually enabled overclocking, increased DRAM voltage to 1.45, changed mhz to 3200, and then I tweaked the timings.

I physically inspected my RAM stick which said "16-18-18-36".

But in the BIOS I only saw 3 numbers that were getting toggled with XMP:
DRAM CAS# Latency > 16
DRAM RAS# to CAS# Delay > 16
DRAM RAS$ ACT Time > 36

I didn't see 4 fields, just those three. Those were being toggled from auto to the above values by XMP, so I entered them manually without applying XMP.

About to go get a coffee and perhaps some food. Should take an hour or so. If the system is still stable when I get back then that's a good sign. Confirmed in Task Manager that the RAM frequency is 3200.

If things are stable at 1.45v, how long would you suggest I wait before trying a lower voltage? And which increments should I use? 1.45 to 1.44?

Thanks,
Stinger

EDIT: I did some digging on what the BCLK does and found this:


In light of this info I have a STRONG suspicion that the base clock manipulation by the XMP profile might be what was causing system instability. As I said, my ram passed 5.5 hours of Memtest and if I turn XMP off, I get no freezes.

If my system doesn't crash for several hours today, I'm going to try putting the voltage back to 1.35 which is what's specified in the XMP profile.

I don't think I ever tried setting the XMP on and then setting BCLK to auto, so I'm hopeful this is what was causing issues.
 
Last edited:

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
Another update:

I'm 99% sure the issue is resolved now. It's been 7 hours since I manually set the RAM overclock and the PC hasn't frozen. I left it idle a few hours and also watched YouTube and played a full round of BF2042 and a bit of Shadow of the Tomb Raider.

Thanks Lutfij for the help. I think the XMP profile setting the BCLK to "100" instead of leaving it on Auto is what caused the issue.

When I manually input the voltage, frequency, and timings, I just left BCLK to auto and I've no issues.

I had initially bumped the voltage to 1.45v, but for the past 3 hours or so I've had it at 1.35v and it's been fine. I don't think voltage was ever the issue.

I think I'll try to get a stable CPU overclock going in the near future. I had always had poor luck doing this (instability), but I think that's primarily because I always tried to use XMP in conjunction with the CPU overclock, and I just assumed I was doing something wrong to the CPU when actually the XMP was screwing me over the whole time.

Thanks again!

-Stinger
 

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
Turns out I spoke a little too soon. I woke up today at 7AM and saw that the PC froze at 4:58AM. So all told, the system was stable for about 20 hours, which is much longer than it had been stable this past week. Of note, it never crashed when I was running the RAM voltage at 1.45. So I've bumped it up again from 1.35 to 1.45 and I've increased the VCCIO and VCCSA to 1.20v . Going to see if I can make it through 24 hours with no freezes.
 
Reactions: deryoman

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
One final update to this thread: The problem was Windows. I had been running Windows 11. The DISM, Scannow, and CHKDSK commands never came up with anything wrong, nor did 5.5 hours of Memtest or a round of Windows' native memory checker. But I finally broke yesterday after another overnight freeze (when I had XMP disabled) and reinstalled Windows. Stable ever since. So my initial hypothesis that it was the RAM or the XMP profile was wrong. I've been running the 3200mhz XMPII profile since about noon yesterday with zero issues. No freezes in over 24 hours, played some rounds of BF2042 and Shadow of the Tomb Raider.

Windows must have gotten corrupted somehow, even though I wasn't getting any obvious indicators that it was.

My only thought on this is that yesterday after I upgraded to Windows 11, I went to reinstall Irfanview (image-viewing program) and afterward, the contextual right-click "Always open this file type with x program" tickbox was gone. Did some digging and found out this program modifies the Windows registry when it reassociates image file types to itself. So I did a system restore to 4 hours earlier (like right after I'd upgraded to Windows 11) which restored the registry and got the program off my machine.

All I can think is that maybe this program or some other made minor tweaks to my registry which just compounded over time and resulted in instability.

Anyway, TLDR; reinstall of Windows solved the problem. I reinstalled Windows 10 from a USB flash drive and then upgraded to Windows 11.
 

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
Minor update: The freezes had started happening again so I did another reinstall of Windows. This time I did a clean install of Windows 11 (I wrote the ISO to a USB stick and everything).

Things were great for about 4 days.

Every time I've reinstalled Windows the freezes cease for days when before they were happening every 3-6 hours. This indicated to me it was a Windows or driver issue (not hardware).

Then yesterday I had a single freeze.

After this I did a hard shutdown and the following:

- Checked for Windows updates. I had a big cumulative one pending which required restart.

- Uninstalled Realtek audio drivers. After the crash I noticed my audio device had defaulted to Sonic Studio 3. I have a Soundblaster AE7 dedicated card. I couldn't just uninstall Sonic Studio but uninstalling Realtek drivers got rid of it.

- Downloaded and installed latest drivers from Asus Armoury Crate utility.

- Downloaded and installed latest Z490-E chipset driver from the mobo support page.

- Downloaded and installed Intel Management Engine.

-Disabled fast boot

Thus far it has been 26 hours with no freezes. So I think one of those things fixed the issue.
 
Last edited:

Stinger86

Distinguished
Feb 23, 2013
15
1
18,515
0
One more (hopefully final) update. I ended up freezing again and decided to replace the motherboard. Replaced my Z490-E with a ROG Maximus XIII Hero. I did this on Feb 2nd and it's Feb 12th now, so 10 days. Zero freezes so far. So I think this was it. Everything is very, very stable. Only XMP enabled, no other overclocks.
 
Mar 29, 2022
1
0
10
0
One more (hopefully final) update. I ended up freezing again and decided to replace the motherboard. Replaced my Z490-E with a ROG Maximus XIII Hero. I did this on Feb 2nd and it's Feb 12th now, so 10 days. Zero freezes so far. So I think this was it. Everything is very, very stable. Only XMP enabled, no other overclocks.
Hi @Stinger86 , I have the exact same problem, with the same motherboard.

Have you faved any freezes after you replace it?
 

ASK THE COMMUNITY