Weird memory timings with Ryzen 5 5600X + freezing issues?

[Ketxxx]

okay, what about this?

If you're doing something special in the registry to cap the CPUs frequency (which is actually easier to do directly from the UEFI anyway) then something is wrong with that component steps like that should never be necessary. W11 is bad, but its not THAT bad. Unless MS ballsed something up with one of their latest updates.

EDIT: looked things over and saw you were using 4.07 chipset drivers. No way the problem is with the drivers. I'm going to say your problem is the CPU or PSU itself if you have already tried installing to a different SSD.

 

[cst1992]

Started getting massive stutters again.

That should be a straight line.

saw you were using 4.07 chipset drivers

I removed those and now am using the 5.08.02.027 drivers, as they are latest. No CLOCK_WATCHDOG_TIMEOUT yet, but other issues, like the above.

 

[VuurVOS]

I would say ignore the stutters and confirm if the system is stable and doesn't freeze/hang anymore.

 

[cst1992]

I would say ignore the stutters and confirm if the system is stable and doesn't freeze/hang anymore.

That's what I'm thinking as well.
I'm getting the stutters both in and out of games (e.g. it takes a bit of time for a window to open), but yeah - no crash yet. If I get a BSOD - with error code or not - I'll post here.

---

1 Dec 2023 1:27PM UTC:
Getting another BSOD, albeit completely different and this time when running sensor scan on HWInfo64.

This is related to either of two things I did on 30 November (reinstallation of driver or setting LLC to 4) as I have logs of 29 November. This means I can't do sensor logging anymore and have to rely on kernel dump.
I don't think this is related to reinstallation of driver (as I have reinstalled that driver before) so why would the LLC setting cause this? Also, is this related to the slowdown I'm currently seeing?

 

[VuurVOS]

That's what I'm thinking as well.
I'm getting the stutters both in and out of games (e.g. it takes a bit of time for a window to open), but yeah - no crash yet. If I get a BSOD - with error code or not - I'll post here.

---

1 Dec 2023 1:27PM UTC:
Getting another BSOD, albeit completely different and this time when running sensor scan on HWInfo64.
View attachment 323639
This is related to either of two things I did on 30 November (reinstallation of driver or setting LLC to 4) as I have logs of 29 November. This means I can't do sensor logging anymore and have to rely on kernel dump.
I don't think this is related to reinstallation of driver (as I have reinstalled that driver before) so why would the LLC setting cause this? Also, is this related to the slowdown I'm currently seeing?

LLC 4 is not the cause of this. It provides a more stable power delivery to the processor.
Since you are getting BSOD then some critical errors are generated. Can you run a stability test for cpu with OCCT?

If there are no errors found after 3x1hour then do the following:
-Clear CMOS
-Disable "Download & Install ARMOURY CRATE app" in Bios (located under tools)
-Load XMP profile
-Save bios changes
-Clean Windows install
-Only install drivers which are provided by the motherboard and graphics card manufacturer.
-Do not install any third party software

Then see if the system is able to survive for a week without a freeze and BSOD.

 

[cst1992]

@VuurVOS Here's one of the traditional freezes that we were checking for (game is Portal 2, by the way).

This is the non-zero Kernel Power event - CLOCK_WATCHDOG_TIMEOUT. I was suspecting if this was exclusive to the 4.x chipset driver - now it's been verified that it happens in 5.x as well.
Doing the CPU stability test now.
By the way, I copied my currently changed BIOS settings to a file. Here's all of the changes currently applied:

Ai Overclock Tuner: Auto -> DOCP
Memory Frequency: Auto -> DDR4-3600MHz
CPU Core Ratio: Auto -> 37.00
PBO Fmax Enhancer: Auto -> Disabled
DRAM CAS# Latency: Auto -> 18
Trcdrd: Auto -> 22
Trcdwr: Auto -> 22
DRAM RAS# PRE Time: Auto -> 22
DRAM RAS# ACT Time: Auto -> 42
Cmd2T: Auto -> 2T
Gear Down Mode: Auto -> Disabled
VDDCR CPU Load Line Calibration: Auto -> Level 4
DRAM Voltage: Auto -> 1.35000
Selects TPM Device: Enable Firmware TPM -> Enable Discrete TPM
When system is in working state: All On -> Aura Off
When system is in sleep, hibernate or soft off states All On -> Aura Off
Above 4G Decoding: Disabled -> Enabled
Overclock: Auto -> Enabled
Precision Boost Overdrive: Auto -> Advanced
Global C-state Control: Auto -> Disabled
Power Supply Idle Control: Auto -> Typical Current Idle

If there's anything obviously weird in this, then let me know.

-Clear CMOS
-Disable "Download & Install ARMOURY CRATE app" in Bios (located under tools)
-Load XMP profile
-Save bios changes
-Clean Windows install
-Only install drivers which are provided by the motherboard and graphics card manufacturer.
-Do not install any third party software

I'll do this if OCCT test is clean.

---

OCCT tests passed. So the CPU is not at fault, at least for the moment.

I'll do another reinstall - this slowdown issue doesn't seem like it'll be easily fixed anyway.

This is looking more and more like a driver thing. No memory, no CPU, no anything else, and now this:

This occurred during shutdown.

 

[VuurVOS]

I'll do another reinstall - this slowdown issue doesn't seem like it'll be easily fixed anyway.

Dont forget to disable "Download & Install ARMOURY CRATE app" in bios before reinstalling Windows.

 

[Ketxxx]

I removed those and now am using the 5.08.02.027 drivers, as they are latest. No CLOCK_WATCHDOG_TIMEOUT yet, but other issues, like the above.

As I said, your issue isn't with the chipset drivers. If you are using more than one pagefile have you disabled them all, rebooted, deleted the pagefile.sys file on the drives if necessary (Windows stupidly doesn't always delete them and to see them you'll need to go to Control Panel > File Explorer Options > View > check the "Show hidden files, folders and drives" option) then enable a pagefile for just one drive? The PAGE_FAULT_IN_NONPAGED_AREA BSOD is typically related to RAM but as there's no way your issue is with RAM it does mean you can rule that out. What can't be ruled out though is potentially a damaged pin in one of the DIMM slots, possibly still a bad IMC, or PSU issue. The more I look at these random errors and that they all seem to happen when the system is under some kind of load, even a small one, the more I think it's worth plugging in an alternative PSU and really taking a close look at the CPU itself, the socket, and the mounting of the cooler. Another thing that comes to mind to try as well is swap the M.2 slot your boot drive is in as if that clears things up we then know either your CPU has an issue, or the M.2 slot has a compatibility problem of some kind with your SSD. Some pictures of the innards of your system would be useful too for sake of being more familiar with your setup if nothing else.

Selects TPM Device: Enable Firmware TPM -> Enable Discrete TPM
Overclock: Auto -> Enabled
Precision Boost Overdrive: Auto -> Advanced

Disable these. If you haven't installed W11 with RUFUS disabling checks for things like the TPM you'll need to make some changes to the registry before disabling TPM. These files I put together should do the job.

On a side note, seriously, TPU STILL doesn't support the .7z format after all these years?

 

[cst1992]

The Overclock thing, I don't use for PBO. All the three values for TDC, EDC and PPT are set to Auto.
I think you have to go to the Overclock section to set C-state Control to Disabled (or something similar, I'll have to check) that's why that is automatically set to Enabled.
I'll reset these settings as @VuurVOS has suggested before the next reinstall.

Disable these. If you haven't installed W11 with RUFUS disabling checks for things like the TPM you'll need to make some changes to the registry before disabling TPM. These files I put together should do the job.

I always use Rufus. No OOBE headache that way, and I've bypassed the TPM stuff while creating the image, too.

On a side note, seriously, TPU STILL doesn't support the .7z format after all these years?

It doesn't support the .log format either. I had to rename my log to .txt for it to be attached.
BTW, do you know how to create .reg files? I want to export some of my changes (like the InprocServer32 thing) for the next time I install Windows.

have you enabled a pagefile for just one drive?

Oh yeah, does this make a difference? I enabled 33GB pagefiles on both my drives after enabling the "complete dump" option to try and make it work. A bit excessive maybe, but yeah.

 

[VuurVOS]

Oh yeah, does this make a difference? I enabled 33GB pagefiles on both my drives after enabling the "complete dump" option to try and make it work. A bit excessive maybe, but yeah.

Dont modify the swap/pagefile and keep it on defaults. Dont enable it on a different drive. Keep the operating system as vanilla as possible without any tweaks.

 

[cst1992]

Did a clean install. CMOS settings applied are as follows:

Ai Overclock Tuner: Auto -> Manual
SB Clock Spread Spectrum: Auto -> Disabled
Performance Enhancer: Auto -> Default (no option to disable)
Memory Frequency: Auto -> DDR4-3600MHz
FCLK Frequency: Auto -> 1800MHz
PBO Fmax Enhancer: Auto -> Disabled
DRAM CAS# Latency: Auto -> 18
Trcdrd: Auto -> 22
Trcdwr: Auto -> 22
DRAM RAS# PRE Time: Auto -> 22
DRAM RAS# ACT Time: Auto -> 42
TRC: Auto -> 68
DRAM Voltage: Auto -> 1.35000
When system is in working state: All On -> Aura Off
Above 4G Decoding: Disabled -> Enabled
Resize BAR Support: Disabled -> Auto
Download and install ARMOURY CRATE App: Enabled -> Disabled

I don't think I will be getting the PAGE_FAULT_IN_NONPAGED_AREA BSODs now - they seem to only occur if the system is especially unstable.
By the way, the Bluetooth adapter is down again, so I am unable to install the driver:

I could no doubt drain capacitors again and install the driver for this adapter, but is this a motherboard fault or device fault?

 

[VuurVOS]

Did a clean install. CMOS settings applied are as follows:

I don't think I will be getting the PAGE_FAULT_IN_NONPAGED_AREA BSODs now - they seem to only occur if the system is especially unstable.
By the way, the Bluetooth adapter is down again, so I am unable to install the driver:

View attachment 323829
I could no doubt drain capacitors again and install the driver for this adapter, but is this a motherboard fault or device fault?

Are you also missing the wifi adapter?

 

[cst1992]

Funnily enough, no. The problem is only in the Bluetooth part of it.

I've held off on installing the driver for it.
The freezes have been happening sooner and sooner, in spite of OS reinstalls. If it doesn't freeze in a good enough timespan this time (a week, say) we'll know it's the card and I'll just change the card out.

 

[Mussels]

View attachment 318920
2 sticks of 3600 MHz RAM - Corsair CMK16GX4M1Z3600C18
Look at the memory timings.
Instead of the optimal 18-22-22-42, I'm getting 26-25-25-58.

My system also tends to experience weird freezing issues - could that be related to this?
Whenever I quit a game or run a memory test, my whole system freezes up such that I can't even open Task Manager. Even Ctrl-Alt-Del doesn't do anything.
I am forced to reset the system using the motherboard's reset button.

System specs:

Processor: Ryzen 5 5600X
Motherboard: Asus ROG Strix X570-E Gaming Wifi II
Cooling: Noctua NH-U12S chromax.black
Memory: 32GB: 2x16GB Corsair Vengeance 3600MHz C18
Video Card(s): NVIDIA RTX 3060Ti Founder's Edition
Storage: 500GB 970 Evo Plus NVMe (boot drive), 2TB Crucial MX500
Display(s): AOC C24G1 144Hz 24" 1080p Monitor
Case: Lian Li O11 Dynamic EVO White
Power Supply: Seasonic X-650 Gold PSU (SS-650KM3)
OS: Windows 10 Pro 64-bit

I have the model down of that motherboard, a few tips for you

1. 3600 is not, never was, and never will be guaranteed to work.
2. Load optimised BIOS defaults. Enable XMP then set SoC voltage to 1.15v in the BIOS. In that order.
3. Infinity fabric raises when your ram speed does, and infinity fabric crashes explain your problems. Higher SoC voltage after XMP is enabled is usually the fix, if it's not lower the RAM speed down to 3200 (AFTER XMP is enabled and the SoC voltage is raised - not before, not in any other order. Enabling XMP changes them back)

The wifi device could be physically loose or faulty, it's a common problem. If it's just having it's USB port drop out then that ties back into the IF crashes and SoC voltage again - whether its USB or PCI-E that drops out varies between systems.

As I said, your issue isn't with the chipset drivers. If you are using more than one pagefile have you disabled them all, rebooted, deleted the pagefile.sys file on the drives if necessary (Windows stupidly doesn't always delete them and to see them you'll need to go to Control Panel > File Explorer Options > View > check the "Show hidden files, folders and drives" option) then enable a pagefile for just one drive? The PAGE_FAULT_IN_NONPAGED_AREA BSOD is typically related to RAM but as there's no way your issue is with RAM it does mean you can rule that out. What can't be ruled out though is potentially a damaged pin in one of the DIMM slots, possibly still a bad IMC, or PSU issue. The more I look at these random errors and that they all seem to happen when the system is under some kind of load, even a small one, the more I think it's worth plugging in an alternative PSU and really taking a close look at the CPU itself, the socket, and the mounting of the cooler. Another thing that comes to mind to try as well is swap the M.2 slot your boot drive is in as if that clears things up we then know either your CPU has an issue, or the M.2 slot has a compatibility problem of some kind with your SSD. Some pictures of the innards of your system would be useful too for sake of being more familiar with your setup if nothing else.



Disable these. If you haven't installed W11 with RUFUS disabling checks for things like the TPM you'll need to make some changes to the registry before disabling TPM. These files I put together should do the job.

On a side note, seriously, TPU STILL doesn't support the .7z format after all these years?

Ask in the suggestions forum? Probably an issue with the forum software not supporting it

 

[cst1992]

I have the model down of that motherboard, a few tips for you

1. 3600 is not, never was, and never will be guaranteed to work.
2. Load optimised BIOS defaults. Enable XMP then set SoC voltage to 1.15v in the BIOS. In that order.
3. Infinity fabric raises when your ram speed does, and infinity fabric crashes explain your problems. Higher SoC voltage after XMP is enabled is usually the fix, if it's not lower the RAM speed down to 3200 (AFTER XMP is enabled and the SoC voltage is raised - not before, not in any other order. Enabling XMP changes them back)

That's quite an old post.
For now, I'm going with whatever the profile changes, nothing more, so as to make minimal changes to defaults.

The wifi device could be physically loose or faulty, it's a common problem. If it's just having it's USB port drop out then that ties back into the IF crashes and SoC voltage again - whether its USB or PCI-E that drops out varies between systems.

Faulty, is my guess. Worked flawlessly for a month, then started having issues with Bluetooth. The driver that came from the motherboard page never worked right (I had to disable then re-enable BluetoothUserService in Task Manager) and another driver version worked better, but not flawlessly. And now the card has failed entirely. No Bluetooth devices are located at all.

Are you also missing the wifi adapter?

Funnily enough, no

Just found out the reason for this. The M.2 key-E interface actually allows for 2x one PCI-e lane + one USB 2.0 connection via the port. The card is two boards in one - one Bluetooth over USB, one WiFi over PCI-e.
This video has an excellent explanation:

 

[VuurVOS]

I think it is wise to RMA the motherboard since it is still in warranty and the bluetooth is acting strange.
This behavior can also be caused by the CPU since a PCI-e lane could be not functioning properly. You can try to reseat the processor a couple of times.

 

[cst1992]

This behavior can also be caused by the CPU since a PCI-e lane could be not functioning properly

Same with USB? Every other USB device apart from this is working fine.

PCIe is the WiFi adapter, which is working fine, no issues at all. Yes, on the same card!

 

[VuurVOS]

Same with USB? Every other USB device apart from this is working fine.

PCIe is the WiFi adapter, which is working fine, no issues at all. Yes, on the same card!

The CPU provides 4x USB 3.2 ports to the motherboard. One of those ports is connected to your bluetooth device.

 

[cst1992]

3-day update:

The Bluetooth device started showing again and reinstalled itself - as in, Windows automatically found and installed its drivers, and the device started showing in Device Manager as normal.

However, I have not tried to use any Bluetooth devices, and I also uninstalled the device's drivers. No freezes so far.

The system also passed a 65-hour corecycler test. I intended it to run for 72 hours straight, but accidentally restarted at the 66-hour mark, iteration 108. That should rule the CPU out.
I also checked if it was seated correctly and it was. No bent pins.

I think this thing was about the Bluetooth all along. I would get freezes from anywhere between 2 and 8 hours, and sometimes longer, but now that I think about it for most of the longer intervals (e.g. 24 hours) I was using the wired keyboard and mouse so that I could access the BIOS if I needed to. In normal use I stuck to my Bluetooth keyboard and mouse.

I have ordered an Intel AX210 card online and will keep using the wired peripherals until it's installed. Then we can see if things keep running okay.

@VuurVOS I know you recommended to RMA the motherboard, but I'd rather not take a chance with the RMA process, in case they send a not-new motherboard back. This motherboard is only 3 months old, and it doesn't seem like anything else is wrong with it.

 

[wNotyarD]

3-day update:

The Bluetooth device started showing again and reinstalled itself - as in, Windows automatically found and installed its drivers, and the device started showing in Device Manager as normal.

However, I have not tried to use any Bluetooth devices, and I also uninstalled the device's drivers. No freezes so far.

The system also passed a 65-hour corecycler test. I intended it to run for 72 hours straight, but accidentally restarted at the 66-hour mark, iteration 108. That should rule the CPU out.
I also checked if it was seated correctly and it was. No bent pins.

I think this thing was about the Bluetooth all along. I would get freezes from anywhere between 2 and 8 hours, and sometimes longer, but now that I think about it for most of the longer intervals (e.g. 24 hours) I was using the wired keyboard and mouse so that I could access the BIOS if I needed to. In normal use I stuck to my Bluetooth keyboard and mouse.

I have ordered an Intel AX210 card online and will keep using the wired peripherals until it's installed. Then we can see if things keep running okay.

@VuurVOS I know you recommended to RMA the motherboard, but I'd rather not take a chance with the RMA process, in case they send a not-new motherboard back. This motherboard is only 3 months old, and it doesn't seem like anything else is wrong with it.

If the new bluetooth-wifi combo card fixes everything, GREAT! We all can finally rest.
If not... Well, only your mobo (and its wifi/bt m.2 slot) left. But fingers crossed it's the card itself.

 

[cst1992]

It should be, because only other explanation is that this is a USB issue.
None of the other USB devices (any storage drives, my Lian-Li fan hub controller, peripherals) have ever reported an issue.
Fingers crossed!

 

[cst1992]

I'm using the new card now. Should be all good from now on.

 

[wNotyarD]

Waiting and hoping for the long-term report.

 

[cst1992]

Well, you could always condense everything thats been tried into a single post to save me like an hours read

You just got your wish. I've updated the first post with the conclusions of everything we've discussed in this thread.

Nothing, and I mean nothing IT related, I've failed to get to the bottom of in 25+ years, I'll be damned if your problem defeats me

Did this one defeat you? It certainly did me...

 

[cst1992]

1-week update after new card installation: No bluescreens/freezes so far. I can FINALLY say the issue is resolved.

I should technically wait for one month or more to know for sure, but I feel I don't need to.

Device Manager and Event Viewer have been clear. No exclamation marks or weird fail events to speak of.