New PC keeps crashing

[PCKettle]

My new PC that I've ordered from Digital Storm that cost me over $6000 keeps crashing and it's becoming to be nearly unusable. Seems to keep getting worse and worse from the initial occurrence and happening more often.

 

Playing world of tanks is impossible, it keeps crashing either after every match or even before I enter the match now while changing the modules. I turned on Prime Video app to watch a show there and my PC blue screened and when I turned on the PC, Prime Video was corrupted and I had to reinstall it to get it to work.

 

I've disabled XMP,  I changed the BIOS CPU settings for less performance and more battery efficiency (although I'm on PC) and that didn't effect anything as well (I've since then reset the BIOS to default settings) and I've updated the BIOS (the update is no longer beta so I went with it) to the most recent version. All of the drivers are up to date and the issue persists. I don't really know what to do.

 

I have contacted Digital Storm and opened a support ticket with them (second time) just have to wait for the weekend to pass to get contacted by them. But the first time I've being talking with them over the phone, they weren't able to help me and completely ignoring the errors in event viewer and their mini dump logs and just having me use my PC for hours until a crash occurs, change one thing, and then use the PC until another crash occurs.

 

The errors in event viewer and their mini dump logs point to CPU, saying there were state issues (where the CPU state gets stuck and doesn't change from idle to whatever it needs to next), a double fault issue and hyper-visor.

 

CPU: Intel 19 14900ks

GPU: MSI GeForce RTX 4090

Motherboard. ASUS ROG Maximus z790 Dark Hero

Ram: 64GB DDR5 6000MHz Kingston FURY Beast RGB

 

One thing I noticed, is that it takes a while for my monitor to get a signal when I'm turning on my PC... sometimes it gets a signal and the PC is already at the log in (past all the logos) making it hard to get into the BIOS if I need to. Last time I needed to get into the BIOS I had to connect the monitor to onboard graphics VIA HDMI (I use DisplayPort on the 4090) and that worked (whether the onboard graphics is faster at giving the signal or if it was a coincidence, I don't know). The monitor I'm using is a Samsung Oddysey G5.

[SpookyCitrus]

What are your CPU and GPU temps, Idle and under load? 

 

Is this one of their custom liquid cooled systems? If so when you filled the loop did you fully bleed the air bubbles out?

 

Revert your bios settings to default, don't do any overclocking or underclocking just yet. Leave XMP disabled for now.

 

Can you list the actual errors your bluescreens have been listing, you can use a program such as bluescreenview to see them easily.

[Mumintroll]

I think since it's brand new (and expensive) you should use the warranty.

Don't do anything that the reseller can find an excuse to say that the warranty is void.

 

I wonder if the reseller has put in a degraded already used cpu and sold as new....

[PCKettle]

Just now, SpookyCitrus said:

What are your CPU and GPU temps, Idle and under load? 

 

Is this one of their custom liquid cooled systems? If so when you filled the loop did you fully bleed the air bubbles out?

 

Revert your bios settings to default, don't do any overclocking or underclocking just yet. Leave XMP disabled for now.

 

Can you list the actual errors your bluescreens have been listing, you can use a program such as bluescreenview to see them easily.

My temps are low, max I saw GPU while playing a game at ultra settings was 76 degrees with the CPU getting around 15% utilization (nvidea experience app doesn't show CPU temps).

 

I hadn't turned on XMP since turning it off and my BIOS is at their default settings (only thing I've changed is changed the logo to display for 5 seconds instead of 3 to help me get into BIOS easier). I'll check on the errors and post them here

[PDifolco]

9 minutes ago, PCKettle said:

My new PC that I've ordered from Digital Storm that cost me over $6000 keeps crashing and it's becoming to be nearly unusable. Seems to keep getting worse and worse from the initial occurrence and happening more often.

 

Playing world of tanks is impossible, it keeps crashing either after every match or even before I enter the match now while changing the modules. I turned on Prime Video app to watch a show there and my PC blue screened and when I turned on the PC, Prime Video was corrupted and I had to reinstall it to get it to work.

 

I've disabled XMP,  I changed the BIOS CPU settings for less performance and more battery efficiency (although I'm on PC) and that didn't effect anything as well (I've since then reset the BIOS to default settings) and I've updated the BIOS (the update is no longer beta so I went with it) to the most recent version. All of the drivers are up to date and the issue persists. I don't really know what to do.

 

I have contacted Digital Storm and opened a support ticket with them (second time) just have to wait for the weekend to pass to get contacted by them. But the first time I've being talking with them over the phone, they weren't able to help me and completely ignoring the errors in event viewer and their mini dump logs and just having me use my PC for hours until a crash occurs, change one thing, and then use the PC until another crash occurs.

 

The errors in event viewer and their mini dump logs point to CPU, saying there were state issues (where the CPU state gets stuck and doesn't change from idle to whatever it needs to next), a double fault issue and hyper-visor.

 

CPU: Intel 19 14900ks

GPU: MSI GeForce RTX 4090

Motherboard. ASUS ROG Maximus z790 Dark Hero

Ram: 64GB DDR5 6000MHz Kingston FURY Beast RGB

 

One thing I noticed, is that it takes a while for my monitor to get a signal when I'm turning on my PC... sometimes it gets a signal and the PC is already at the log in (past all the logos) making it hard to get into the BIOS if I need to. Last time I needed to get into the BIOS I had to connect the monitor to onboard graphics VIA HDMI (I use DisplayPort on the 4090) and that worked (whether the onboard graphics is faster at giving the signal or if it was a coincidence, I don't know). The monitor I'm using is a Samsung Oddysey G5.

14900KS ? Well it surely has degraded quickly...

RMA it

[PCKettle]

2 minutes ago, Mumintroll said:

I think since it's brand new (and expensive) you should use the warranty.

Don't do anything that the reseller can find an excuse to say that the warranty is void.

 

I wonder if the reseller has put in a degraded already used cpu and sold as new....

I did open a support ticket with them a second time, but I lost any faith/trust I had in them after the first support ticket. Although the issue was sparse then and it seemed like the issue was solved after a few hours, but it didn't help that they ignored all the logs/errors in event viewer, saying that they are irrelevant.

[SpookyCitrus]

2 minutes ago, PCKettle said:

My temps are low, max I saw GPU while playing a game at ultra settings was 76 degrees with the CPU getting around 15% utilization (nvidea experience app doesn't show CPU temps).

 

I hadn't turned on XMP since turning it off and my BIOS is at their default settings (only thing I've changed is changed the logo to display for 5 seconds instead of 3 to help me get into BIOS easier). I'll check on the errors and post them here

Use HWMonitor to view your CPU temps at both idle and under load. 

[PCKettle]

Oh, and it's not liquid cooled. I use an AiO for the CPU and that's about it. I'm using the Aventum X case from Digital Storm and it seems to have plenty of airflow

[SpookyCitrus]

1 minute ago, PDifolco said:

14900KS ? Well it surely has degraded quickly...

RMA it

I highly doubt Digital Storm didn't check their stock for degradation, nor that they would ship a system out without the microcode updates. Unless OP's new system they just ordered is actually one they've had for several months already.

[PCKettle]

1 minute ago, SpookyCitrus said:

I highly doubt Digital Storm didn't check their stock for degradation, nor that they would ship a system out without the microcode updates. Unless OP's new system they just ordered is actually one they've had for several months already.

I got the PC mid September, the new BIOS with the most recent intel micropatch came out in October. I've updated the BIOS already and that didn't fix the issue. Digital Storm had the BIOS at the August 8 update (which was the most recent at the time I got it).

[SpookyCitrus]

Just now, PCKettle said:

I got the PC mid September, the new BIOS with the most recent intel micropatch came out in October. I've updated the BIOS already and that didn't fix the issue. Digital Storm had the BIOS at the August 8 update (which was the most recent at the time I got it).

The August 8th update it had should have been the initial microcode update 0x129, so degradation shouldn't be a factor here unless the CPU was previously used(which I doubt, but isn't impossible). I'd really like to know your CPU temps, and what specific BSOD errors it's giving you. Let us know once you test them. If you don't already have it Cinebench is a good option for stability and stressing the CPU to view temps and check for crashing, so I'd monitor them with HWMonitor while you run Cinebench, you'll know pretty quickly if the CPU is overheating while stressed.

[PCKettle]

The bluescreenview didn't show any errors but before my PC blue screened, there is a warning with the following contents, "The hypervisor did not enable mitigations for side channel vulnerabilities for virtual machines because HyperThreading is enabled. To enable mitigations for virtual machines, disable HyperThreading." and the BSOD had the HYPERVISOR_ code (don't remember what was after the line)

 

A previous BSOD had the stop code as 'CLOCK_WATCHDOG_TIMEOUT'

[PDifolco]

21 minutes ago, SpookyCitrus said:

I highly doubt Digital Storm didn't check their stock for degradation, nor that they would ship a system out without the microcode updates. Unless OP's new system they just ordered is actually one they've had for several months already.

Wasn't saying that, OP don't tell since when he has the PC, was assuming some months

Even Intel is willing to replace failing 14th gen chips

[Mumintroll]

17 minutes ago, PCKettle said:

The bluescreenview didn't show any errors but before my PC blue screened, there is a warning with the following contents, "The hypervisor did not enable mitigations for side channel vulnerabilities for virtual machines because HyperThreading is enabled. To enable mitigations for virtual machines, disable HyperThreading." and the BSOD had the HYPERVISOR_ code (don't remember what was after the line)

 

A previous BSOD had the stop code as 'CLOCK_WATCHDOG_TIMEOUT'

 

CLOCK_WATCHDOG_TIMEOUT could point to a bad cpu or perhaps something with the memory.

 

[SpookyCitrus]

19 minutes ago, PCKettle said:

The bluescreenview didn't show any errors but before my PC blue screened, there is a warning with the following contents, "The hypervisor did not enable mitigations for side channel vulnerabilities for virtual machines because HyperThreading is enabled. To enable mitigations for virtual machines, disable HyperThreading." and the BSOD had the HYPERVISOR_ code (don't remember what was after the line)

 

A previous BSOD had the stop code as 'CLOCK_WATCHDOG_TIMEOUT'

If you are getting Hypervisor errors related to hyperthreading you can try enabling the core scheduler.

 

Open CMD in administrator and run the following command.

 

"bcdedit /set hypervisorschedulertype core" (without the quotes).

 

Exit CMD and use the computer like normal, do things that were causing it to crash previously, the more repeatable the better. If it no longer crashes, that was the issue.

 

If it still crashes the only recommendation I could make is creating a Windows 11 installer on a USB drive and format and reload Windows 11. That will guarantee it is not a software issue. With teh CLOCK WATCHDOG TIMEOUT error it very well could be a bad CPU, re-installing Windows 11 will confirm it's not some weird software issue. And then you'll know your next step will be replacing the CPU via Intel RMA or contacting DS for a replacement. 

[PCKettle]

I ran cinemabench with HWMonitor alongside to keep tracks of the temperatures, and here is what I've found out.

 

Each test was (supposed to be) 10 minutes.

 

GPU Test ran the entire 10 minutes and completed 60 passes during which the CPU stayed around 50 degrees during a pass, but spiked to 100 between some passes (and sometimes didn't spike all the way to 100, sometimes to 70). By CPU, I mean under the 'package' portion of HWMonitor for CPU. P-Cores also spiked to 100 between passes sometimes. E-Cores stayed 45 degrees throughout and GPU stayed around 65 the entire time (varying between 57 and 70; but stuck mostly around 65). Finished with 32758 points.

 

CPU Multi-core test I let run for a little over 1 pass (about 3 minutes) since as soon as I started the test the CPU got to 100 degrees and stayed there for the entire time. After the first pass and before the second pass (when it was preparing the project) the CPU temp fell down to 70 before jumping back up to 100 at the second pass (I stopped the test as soon as the second pass started). P-Core also ran at 100 degrees the entire time, dropping between passes and I didn't record the E-Core temperature here. GPU stayed around 43 degrees the whole time. I didn't take notice of the running points score during this test.

 

CPU Single-core test ran for 9 minutes and 40 seconds out of the 10 minutes it was supposed to run and froze entirely with 20 seconds left to run. Cinemabench was the only thing that froze while HWMonitor kept working and as soon as Cinemabench froze, the CPU temps dropped to around 40 degrees. While it was running though, the CPU package temps stayed around 70 degrees sometimes getting into the 80s. P-Core stayed mostly around 60s and sometimes getting into the 80s. E-Core stayed consistently in the 40s and the GPU stayed at about 48 the entire time. It was at around 120 points when it crashed, took 8 minutes to get to 120 points, before that, it stayed around 90-107 points.

 

After I restarted cinemabench, I ran the CPU multi-core test again with the same results as the first run, but this time I recorded the e-core temps at 90-93 degrees the entire time (93 degrees being the max according to HWMonitor). This time around, I checked the running points during the pass and it was around 2020 for the whole duration of the pass.

 

I stopped the tests for multi-core tests myself after 1 pass because the CPU package temps stayed at 100 the entire times... didn't want to make things worse.

[PCKettle]

21 hours ago, SpookyCitrus said:

If you are getting Hypervisor errors related to hyperthreading you can try enabling the core scheduler.

 

Open CMD in administrator and run the following command.

 

"bcdedit /set hypervisorschedulertype core" (without the quotes).

 

Exit CMD and use the computer like normal, do things that were causing it to crash previously, the more repeatable the better. If it no longer crashes, that was the issue.

 

If it still crashes the only recommendation I could make is creating a Windows 11 installer on a USB drive and format and reload Windows 11. That will guarantee it is not a software issue. With teh CLOCK WATCHDOG TIMEOUT error it very well could be a bad CPU, re-installing Windows 11 will confirm it's not some weird software issue. And then you'll know your next step will be replacing the CPU via Intel RMA or contacting DS for a replacement. 

Enabling the core scheduler didn't fix the issue, nor did reinstalling Windows. Furthermore, I noticed now (after re-install) Windows doesn't seem to remember desktop view settings (I don't want icons to align to grid). This isn't a main issue, but once the CPU issue is fixed, this will become irritable.