New Ryzen build - system freezes - not the a-typical black screen or WHEA

[newconroer]

After some great advisement by fellow members, I have recently platform JUMPED (ok that might be extreme) from Sandybridge to AM5
And while it's pretty much been good overall, I am stumped about system crashing that is occurring

I might have to update this post as I go as I will invariably think of something later that was missed out



Symptom
System freezes and stays that way indefinitely until I restart manually

• The audio will keep playing (though it does not stutter and lead onto a BSOD
• Visuals are frozen
• Light on keyboard and other peripherals go out
• Input of any type has no effect

Observations

Only seems to occurs when the following is true :

• -3D Application such as a game is running

and

• - A web browser is open

Only seems to occur when the above is true and I believe also (anecdotally) that :

• I haven't been using the system for a moment or
• I alt tabbed or moved mouse to other monitor or
• I do a scroll on the browser on other monitor
  E.g. under load it doesn't occur - and by load I mean the 3D program/game is the focus window(not necessarily full load as in boosting cores/maximum output)

Other :

• When it happens I hear a slight jump in fan speed which suggests to me the CPU has been ramped up/possibly boosting and therefore outputting more heat - of note it's not excessively high, only a moderate increase in fan cycle, which to me (since I set a curve) I know is still within safe temperatures. May not be relevant though might give weight to the idea this could be something to due with sleeping or parked cores trying to spring into action then crashing
• It never occurs during boot or start up of windows (login, services start, programs start etc).
• Some occasions ( not the majority), there's activity even after it freezes - where the screen ever so slightly flickers. By flicker I mean a pulsing of brightness higher, then lower, then higher etc. This might be the physical screen refreshing (yet the supplied image is static), causing some type of anomaly our eyes can detect.
• The time span or interval(s) before it begins is dependent upon when I start a 3D program + a browser e.g. I could go for quite a while without the issue, and then when I meet above conditions it can happen five minutes, or forty-five minutes later
• Memory dumps are infrequent, there isn't one every time or even every other time that this occurs(and the ones I have seen are not very helpful)
• After each crash, my mouse settings through Razer Synapse do not work. The program/app loads, however I have to manually exit and relaunch it. Then it is OK
  This happens which each crash/manual reset
• Can occur when running 3D application or game from two different drives. e.g. One NVME SSD and the other a rust disk

Windows event viewer /error tracking :

• Windows event viewer has no clear errors that shed light on this or seem related - I can see the event log entries and Kernel entries about unplanned system shutdown though not what caused it
• There's only three WHEA errors and those only occurred a few weeks ago when I was doing some manual voltage testing and was too undervolted to be stable
• I do have some critical and warnings (mostly warnings) entries however :
  A) The timing of them does not always align with the crashes or anywhere remotely close
  B) The frequency of them is such that it happens multiple times in between each crash. One would think if it was causing the crash it wouldn't error multiple times THEN cause the crash, it would be the first time every time
  
  Examples include
  
  Critical
  Event 131, DeviceSetupManager - "Metadata staging failed, result=0x80004005 for container '{2D4CAE54-5B69-5328-BC5D-AF10B6CDF8DF}'"
  Event 10010, Distributed Com - "The server {8CFC164F-4BE5-4FDD-94E9-E2AF73ED4A19} did not register with DCOM within the required timeout."
  
  Warning
  Event 122, DeviceSetupManager - "Access to drivers on Windows Update was blocked by policy"
  Event 200, DeviceSetupManager - "A connection to the Windows Update service could not be established."
  Event 201, DeviceSetupManager - "A connection to the Windows Metadata and Internet Services (WMIS) could not be established."
  Event 202, DeviceSetup Manager - "The Network List Manager reports no connectivity to the internet."
  Event 360, User Device Recognition - "Windows Hello for Business provisioning will not be launched. Device is AAD joined ( AADJ or DJ++ ): Not Tested
  User has logged on with AAD credentials: No
  Windows Hello for Business policy is enabled: Not Tested
  Windows Hello for Business post-logon provisioning is enabled: Not Tested
  
  Event 10016, Distributed Com - "The application-specific permission settings do not grant Local Launch permission for the COM Server application with CLSID
  Windows.SecurityCenter.WscDataProtection
  and APPID Unavailable
  to the user NT AUTHORITY\SYSTEM SID (S-1-5-18) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool."

Things I have tried


System/Applications/Services

• All BIOS settings to default
• Latest BIOS from manufacturer
• All motherboard recommended standard drivers(Chipset, LAN, etc.)
• No devices outstanding in Device Manager
• Latest Windows update
• Starting Windows with selective - Microsoft only /non user services
• Starting Windows with virtually no programs /user programs
• Used both Edge and Brave browser (relevance here is it seems to only happen when browser is running)
• Disable/enable hardware acceleration (for example had off in Edge and on in Brave)
• Windows power plan with minimum processor state at 5% versus 90% versus 95%
• Power plan adjusted to not allow anything to auto turn off or goto sleep or similar
• Amended BIOS for testing with :
  • PSU idle power behavior to "typical" instead of "auto" or "low"
  • Disabled C-states and DF states
  • Attempted to set positive offset voltage( no joke, I cannot - apparently according to AsRock this was disabled, yet you can still change voltages overall...)
  • Set the DDR Power management to disable the idle state /lower power state of DDR5
  • Set RAM to use EXPO to see whether it would bump the voltage as part of the settings to give more stability(assuming the automatically supplied voltage is not enough)
  • Set RAM manual timings to ensure they match the advertised product speeds

Graphics /Video card/hardware

• Windows graphics settings (GPU acceleration, Variable Rate, Games Optimization disabled)
• Graphics drivers removed reinstalled
• Disabled onboard GPU in Device manager
• Disconnecting any additional monitors physically (or in software /both)
• Tried to recreate while having Edge/Brave on the main monitor versus a secondary or tertiary monitor - happens on both monitors
• Reseated GPU in motherboard
• Checked mount studs for case - > motherboard they seem OK
• Set Nvidia properties to allow all programs, even browser to use Performance /Maximum ; also tested with Optimal Power setting

Miscellaneous

• Loaded up a Windows install from another drive
• Checked temperatures for CPU, Cores, motherboard components/VRM, RAM, SSD/rust drives, GPU
• Power delivery seems OK - I find no power bottlenecking /delivery issues and when programs are going, they are going without issue (clock speeds boost, wattage is delivered as expected)

Tests run
(no errors found)

• Ycruncher, Prime 95 and Core cycler
• OCCT
• Memtest64and classic RAM test
• AIDA 64 benchmarks
• Cinebench
• Various game or 3Dmark benchmarks

=--=


Thoughts thus far..

• Given it happens on stock settings that is concerning
• Hardware cause instead of software seems more likely be the case as it also occurs when I load up the other Windows system install on a separate SSD drive which admittedly is also Windows 11 though a previous version and - of note doesn't included all the same drivers, as it was from the Sandybridge system prior
• It seems to be more on a low power or low clock usage, where maybe coming out of sleep state or similar is resulting in a crash. And I could understand when using a heavy negative Curve Optimizer as it is known to be unstable at low voltage draw or frequency. However Curve is off and I also disabled the C and DF states in the testing
• It's not the a-typical symptoms you would associate with unstable system due to low voltage setting or in Ryzen's case PBO and Curve Optimizer (even if there was one) i.e. no BSOD, no WHEA, no black screen and system reboot
• I don't believe the aforementioned Event Viewer entries are related. Reading around they seem to be par for the course with Windows and often benign. As well a red herring or misleading - for example it states my Network List is not working or indicates I have no network connection however I do
• Memory dumps haven't been proved to be much use though should anyone else want a copy to take a crack at looking over it, please ask
• I am in possession of a system restore that is /was made not long after Windows install, with some drivers updated and effectively a baseline before any custom/user stuff was installed. I have not tried restoring to that point yet
• Real time note : I have been web browsing, doing this post and etc on/off for forty minutes - no issues,

Trying to think of some other tests or ways to do analysis to get at least some idea of where this is stemming from just to get a baseline
I wish this was a simple as straight forward as an extreme negative Curve Optimizer was resulting in black screens or WHEA - at least then it makes sense

 

[ir_cow]

Does this do it to you only under load or at idle as well? - The black screen.

I've experienced this before and it has been for me (at least), video card once and a SSD another time related. Many more times over the years, but that's all I can think of that is most recent.

Just a few months ago it was the SSD, all games would eventually freeze 10 minutes to 3 hours. Until I decided to move games to a different drive and all the problems went away. I went down the list until I was out of ideas lol.

 

[RJARRRPCGP]

Those event log messages are Windows bugs. Except for the Event 122 warning, because of Windows Update being blocked from installing drivers. (normal)

 

[newconroer]

It is definitely more on the lower cycle/idle side
Keep in mind I am not black screening, it's freezing yet all visuals are still on screen

I am mindful about the GPU however it was the same one I had in previous build where this behavior did not exist.

I'll take your point about the hard drive though I updated the thread now to comment on that thank you

Those event log messages are Windows bugs. Except for the Event 122 warning, because of Windows Update being blocked from installing drivers. (normal)

Appreciate the feedback

 

[R0H1T]

Did you try turning off hardware acceleration for the web browser? This is a major issue with chromium based browsers, at least for me & even for the same version (111.11.1x et al) one browser could have this issue like Chrome & the other may not like Opera. Which is really frustrating because even if you disable any feature flags the issue could still persist, even with just default configuration & only hardware acceleration tuned on! I've had to manually tune many browsers like that & because I use more than one it's a real PITA

 

[newconroer]

Did you try turning off hardware acceleration for the web browser? This is a major issue with chromium based browsers, at least for me & even for the same version (111.11.1x et al) one browser could have this issue like Chrome & the other may not like Opera. Which is really frustrating because even if you disable any feature flags the issue could still persist, even with just default configuration & only hardware acceleration tuned on! I've had to manually tune many browsers like that & because I use more than one it's a real PITA

Aye, I have
Updated post above to include this now

And it may not be the browser, just seems that whenever this happens it is open versus when it's not open I cannot recall this happening
Though still going to try again this weekend for another long stint of several hours with no browsers open

 

[mechtech]

Sometimes an order of operations matters.......or at least helps..........

when I do a new build I usually

set bios to default
confirm xmp., etc. is Disabled in bios
confirm UEFI in bios enabled
I make a memtest on a bootable USB and let it run 24 hours before installing windows
windows iso - i will confirm checksum before making bootable usb - usually check usb drive to make sure it doesn't have bad sectors, etc.
make sure GPT when formatting/installing windows (usually disconnected from internet)
install chipset driver
run some tests on a clean windows install with no internet to make sure it's good, check CPUz, etc., to confirm voltages, frequencies etc.

if its all good after that - then make a restore point and start installing things, one at a time with a reboot between each one to make trouble shooting easier if something goes south

 

[Super Firm Tofu]

I'm going to say it's a problem with the older design PSU. That model came out in like 2011. Do you have a newer unit you can try?

 

[newconroer]

Sometimes an order of operations matters.......or at least helps..........

when I do a new build I usually

set bios to default
confirm xmp., etc. is Disabled in bios
confirm UEFI in bios enabled
I make a memtest on a bootable USB and let it run 24 hours before installing windows
windows iso - i will confirm checksum before making bootable usb - usually check usb drive to make sure it doesn't have bad sectors, etc.
make sure GPT when formatting/installing windows (usually disconnected from internet)
install chipset driver
run some tests on a clean windows install with no internet to make sure it's good, check CPUz, etc., to confirm voltages, frequencies etc.

if its all good after that - then make a restore point and start installing things, one at a time with a reboot between each one to make trouble shooting easier if something goes south

Hear you on that, I did exactly that - got installed to a reasonable 'baseline'

I haven't restored to that date though - will mention in original post
Should I make a restore point now, will windows allow me to go back then forward again ?

I'm going to say it's a problem with the older design PSU. That model came out in like 2011. Do you have a newer unit you can try?

I have another PSU however it's older
I am also mindful it was fine in the Sandybridge system a month ago so it hasn't been at top suspects list

Should I buy one if nothing else as a test- any recommendations? I am not up on current Power Supply technology
I take the argument over single high 12amp rail versus multi rail is a dead one
And that 80 Platinum is the new 'gold' standard

I would probably get an 850 watt again, something that is modular for sure

 

[eidairaman1]

Seasonic or Superflower or Andyson Based

 

[AlwaysHope]

OP, you've upgraded to an AM5 system & your using a decade old PSU... come on man, get real here! that PSU may seem fine in older systems but newer ones not so much.
Seriously, upgrade that PSU asap. A gold standard grade is perfectly fine, no need to go overboard unless you have $ to burn.

 

[newconroer]

Seasonic or Superflower or Andyson Based

Thanks

One thing I didn't mention in my post (although I cannot modify /edit any posts now today....) is that I BIOS flashed successfully to their 1.21 which
was created to address the SoC voltage issue on AMD motherboards


1. Update AGESA ComboAM5 1.0.0.6.
2. Improve memory compatibility.
3. Optimize for Ryzen 7000X3D series processor.
4. Recommend for upgrade BIOS if 7000X3D processor is installed.

There's another one out however seems reports are that's a poor BIOS causing lots of issues
The next one for AGESA 1.0.0.8 (or .9) is in testing currently. I am hoping that is out soon and I can move to that version

 

[dgianstefani]

OP, you've upgraded to an AM5 system & your using a decade old PSU... come on man, get real here! that PSU may seem fine in older systems but newer ones not so much.
Seriously, upgrade that PSU asap. A gold standard grade is perfectly fine, no need to go overboard unless you have $ to burn.

Absolutely. Power supplies can still "work", while having degraded output due to age and wear. The specs it had at new are certainly not what it will be running now.

Check out our reviews for PSUs https://www.techpowerup.com/review/...ditorschoice=1&recommended=1&pp=25&order=date.

https://cultists.network/140/psu-tier-list/ This list is also somewhat reliable if you choose a unit from a higher tier.

You don't need to overspend, as your system is relatively low powered, but I'd look into getting a good quality 850 W + platform, to somewhat future proof you, and as you seem to keep PSUs for a long time. Corsair/Seasonic would be my first choice, but there are other good manufacturers too.

If I were planning to get a PSU I'd keep for a decade it would be a Seasonic Prime unit, or for SFX, what I have now.

Look into getting an ATX 3.0 PSU if you can afford the extra 20% premium or so, it's a good standard moving forwards.

Even if this turns out to not be the root cause of your main issue, it's still a very good idea to replace a PSU that's a decade old (assuming it's seen regular use).

To troubleshoot, you could remove the GPU and use the iGPU in your 7600X, perhaps the reduced power load would eliminate the issue, but it depends how worn out your PSU is.

 

[TheoneandonlyMrK]

Those event log messages are Windows bugs. Except for the Event 122 warning, because of Windows Update being blocked from installing drivers. (normal)

Indeed I was about to say recently and without any prior issue I had this happen on my Am4 rig after a clean install, the audio continued that had not happened before I actually thought the picture a Yt vid was caching.

Maybe not tied but maybe just OS

 

[Shou Miko]

OP, you've upgraded to an AM5 system & your using a decade old PSU... come on man, get real here! that PSU may seem fine in older systems but newer ones not so much.
Seriously, upgrade that PSU asap. A gold standard grade is perfectly fine, no need to go overboard unless you have $ to burn.

Not gonna tell u how old my Seasonic Prime 1.2K watt is than

 

[dgianstefani]

Not gonna tell u how old my Seasonic Prime 1.2K watt is than

You have a 1.2 KW unit, which is probably still able to output at least 800 W, even with degradation. An 850 W unit may have degraded to 600 W or even less.

 

[Shou Miko]

You have a 1.2 KW unit, which is probably still able to output at least 800 W, even with degradation. An 850 W unit may have degraded to 600 W or even less.

Last year it ran a RTX 3090 fin even overclocked so I think it still does more but these days it's having a lighter load if you check my system specs even I just upgraded from a RX 6800 XT Red Devil to a RX 7900 XT Speedster Merc310.

 

[newconroer]

Absolutely. Power supplies can still "work", while having degraded output due to age and wear. The specs it had at new are certainly not what it will be running now.

Check out our reviews for PSUs https://www.techpowerup.com/review/...ditorschoice=1&recommended=1&pp=25&order=date.

https://cultists.network/140/psu-tier-list/ This list is also somewhat reliable if you choose a unit from a higher tier.

You don't need to overspend, as your system is relatively low powered, but I'd look into getting a good quality 850 W + platform, to somewhat future proof you, and as you seem to keep PSUs for a long time. Corsair/Seasonic would be my first choice, but there are other good manufacturers too.

If I were planning to get a PSU I'd keep for a decade it would be a Seasonic Prime unit, or for SFX, what I have now.

Look into getting an ATX 3.0 PSU if you can afford the extra 20% premium or so, it's a good standard moving forwards.

Even if this turns out to not be the root cause of your main issue, it's still a very good idea to replace a PSU that's a decade old (assuming it's seen regular use).

To troubleshoot, you could remove the GPU and use the iGPU in your 7600X, perhaps the reduced power load would eliminate the issue, but it depends how worn out your PSU is.

I like this idea and I did consider it, however the internal GPU may not be able to load some games/3D programs? making it unable to recreate the scenario.
Something I'll have look into again

 

[newconroer]

After some more tedious testing, it seems this is the cause

I thought I had disabled it prior (and likely I did - though probably enabled it again in haste or by accident )