Question Disabling C-States fixed freezes... now what?

[guysbeingdudes]

System Specs
CPU: Ryzen 3950x
Motherboard: TUF GAMING X570-PLUS (WI-FI)
Ram: 2x16GB G.Skill Trident Z Neo Series DDR4 3600
SSD/HDD:
1x Samsung 970 EVO Plus SSD 2TB
1x Samsung 980 SSD 500GB
GPU: GeForce GTX 1080 Ti
PSU: Corsair RMx Series RM850x
OS: Windows 10 on 2TB SSD, Linux Mint on 500GB SSD
-NO overclocking

I've been having issues for the past few months where my system will freeze in place on a screen, forcing me to reset. This happens almost every day, usually on the lock screen. If I then unplug the monitors from the GPU and plug them back in, the screen does not come back up. I am using two monitors, one connected by HDMI and one by DP. I just recently (past month) installed a 2nd SSD to run Linux Mint. I have witnessed this occur on Windows, on Linux, and even in the BIOS screen.

I have witnessed this issue occurring on the lock screen with a clock display, so I am able to pinpoint to the second when the freeze occurred. However, checking the syslog shows no activity or information around the time of freeze.

Fixes I've already attempted with no avail:
-Updated BIOS
-Updated and reverted GPU drivers
-Upgraded DP Firmware for GPU
-Reseat GPU
-Change PS cables to CPU
-Ran memtest overnight to check for RAM issues, 7 passes and no errors

The only thing that solved the issue was Disabling C-States from the BIOs. Since doing this my system has been up for +5 days straight with no problems or overheating.

Now I'm left wondering what the cause of the problem is in the first place and also if I can do anything else to improve stability. The system was built within the past 14 months, and all of the parts are new except the GPU, so RMA is an option for any faulty parts.

Any help or advice is greatly appreciated!

 

[kanewolf]

System Specs
CPU: Ryzen 3950x
Motherboard: TUF GAMING X570-PLUS (WI-FI)
Ram: 2x16GB G.Skill Trident Z Neo Series DDR4 3600
SSD/HDD:
1x Samsung 970 EVO Plus SSD 2TB
1x Samsung 980 SSD 500GB
GPU: GeForce GTX 1080 Ti
PSU: Corsair RMx Series RM850x
OS: Windows 10 on 2TB SSD, Linux Mint on 500GB SSD
-NO overclocking

I've been having issues for the past few months where my system will freeze in place on a screen, forcing me to reset. This happens almost every day, usually on the lock screen. If I then unplug the monitors from the GPU and plug them back in, the screen does not come back up. I am using two monitors, one connected by HDMI and one by DP. I just recently (past month) installed a 2nd SSD to run Linux Mint. I have witnessed this occur on Windows, on Linux, and even in the BIOS screen.

I have witnessed this issue occurring on the lock screen with a clock display, so I am able to pinpoint to the second when the freeze occurred. However, checking the syslog shows no activity or information around the time of freeze.

Fixes I've already attempted with no avail:
-Updated BIOS
-Updated and reverted GPU drivers
-Upgraded DP Firmware for GPU
-Reseat GPU
-Change PS cables to CPU
-Ran memtest overnight to check for RAM issues, 7 passes and no errors

The only thing that solved the issue was Disabling C-States from the BIOs. Since doing this my system has been up for +5 days straight with no problems or overheating.

Now I'm left wondering what the cause of the problem is in the first place and also if I can do anything else to improve stability. The system was built within the past 14 months, and all of the parts are new except the GPU, so RMA is an option for any faulty parts.

Any help or advice is greatly appreciated!

Usually disabling C states causes the CPU to run at a constant, less than max clock speed. I would start looking at cooling or voltage stability.

 

[guysbeingdudes]

Usually disabling C states causes the CPU to run at a constant, less than max clock speed. I would start looking at cooling or voltage stability.

I have been monitoring the temps of the CPU and GPU for the past 5 days as well. They are staying constant at ~40C.

 

[Karadjgne]

C-States are the step downs at idle, the sleep states. C0 is On, active. C1-3 are the first layers of sleep, ones that include the lock screen, screen savers etc where the cpu is basically still active but has shut down non-essential programs and services. C4-5 shuts down almost everything basically except for windows timers and lowers voltages further. C6+ is deep sleep where even windows timers are put into a coma.

The problem is doubtfully the psu, it can handle the voltage reductions, it's more likely a program conflict where a program wants to run, but cannot, so gets stuck in an endless loop of trying to start. This will lock up the pc.

Malware is often the original culprit, there are sleepers in the registration that activate periodically to check to see if the malware is still installed, if it's been removed or deleted, it'll activate a refresh download/install.

The worst part is finding the problem programming because you need to disable the services, let the pc go to sleep, try and wake it up, and keep doing that for any extra programs that auto run, not games, but stuff like added lock screens, update programs, driver updaters etc.

 

[guysbeingdudes]

C-States are the step downs at idle, the sleep states. C0 is On, active. C1-3 are the first layers of sleep, ones that include the lock screen, screen savers etc where the cpu is basically still active but has shut down non-essential programs and services. C4-5 shuts down almost everything basically except for windows timers and lowers voltages further. C6+ is deep sleep where even windows timers are put into a coma.

The problem is doubtfully the psu, it can handle the voltage reductions, it's more likely a program conflict where a program wants to run, but cannot, so gets stuck in an endless loop of trying to start. This will lock up the pc.

Malware is often the original culprit, there are sleepers in the registration that activate periodically to check to see if the malware is still installed, if it's been removed or deleted, it'll activate a refresh download/install.

The worst part is finding the problem programming because you need to disable the services, let the pc go to sleep, try and wake it up, and keep doing that for any extra programs that auto run, not games, but stuff like added lock screens, update programs, driver updaters etc.

Oh man, this sounds like it’ll be absolute hell to debug! Do you have any recommendations on how to handle it?

 

[Karadjgne]

Start out with a good malware screen. A virus screen also wouldn't hurt. Then run through the list of programs you have installed. Some, like Adobe or Photoshop or Java are pretty much universally safe. What you'd be looking for is the not so well known, and not so well coded, like driver updaters.

Then I'd run ccleaner and also the registration tool included, at its default settings (say yes to save a copy). I'd also go to Asus website and update the motherboard specific drivers package.

 

[guysbeingdudes]

Start out with a good malware screen. A virus screen also wouldn't hurt. Then run through the list of programs you have installed. Some, like Adobe or Photoshop or Java are pretty much universally safe. What you'd be looking for is the not so well known, and not so well coded, like driver updaters.

Then I'd run ccleaner and also the registration tool included, at its default settings (say yes to save a copy). I'd also go to Asus website and update the motherboard specific drivers package.

What are your thoughts on the issue occurring across both OS while dual booting from two different SSDs? Could a problem with Windows 10 cause Linux Mint to freeze or vice versa?

 

[Karadjgne]

OS is software. When you start the pc, you'll be running post and bios which has its own drivers etc that it needs in order to make the hardware work. Then the OS takes over, loads up the registry, sorts itself out as far as drivers, startup services, programs etc, go.

So Windows and Linux are seperate, have their own advanced drivers and procedures. But. That's just bios/OS. It's not any programs or drivers installed that operate under either OS. You'd need 2x seperate entirely installations, so the windows does not see anything used by Linux and vice versa. Like haveing 2 seperate installs of Adobe, not 1 installed Adobe used by either.

As long as there is any shared programming, what affects one install can affect the other. Also Windows and Linux (or any of its Distros) are different. It's not uncommon for software or hardware to run butter smooth in Linux and run like 80grit sandpaper in windows, or vice versa.

 

[guysbeingdudes]

OS is software. When you start the pc, you'll be running post and bios which has its own drivers etc that it needs in order to make the hardware work. Then the OS takes over, loads up the registry, sorts itself out as far as drivers, startup services, programs etc, go.

So Windows and Linux are seperate, have their own advanced drivers and procedures. But. That's just bios/OS. It's not any programs or drivers installed that operate under either OS. You'd need 2x seperate entirely installations, so the windows does not see anything used by Linux and vice versa. Like haveing 2 seperate installs of Adobe, not 1 installed Adobe used by either.

As long as there is any shared programming, what affects one install can affect the other. Also Windows and Linux (or any of its Distros) are different. It's not uncommon for software or hardware to run butter smooth in Linux and run like 80grit sandpaper in windows, or vice versa.

I’m wondering how can I be seeing the freezing occur in both OS? Like you said, Windows does not see anything used by Linux and vice versa. So a faulty program on Windows shouldn’t cause freezing on Linux and vice versa correct?

Does the issue occurring across both OS point to the root of the problem likely being hardware related, or am I off base here?

 

[Karadjgne]

Do a test. Unplug 1 ssd data cord, just allow the other OS to load by itself. See if it freezes. Then try the reverse. That'll tell you if it's something that's common to both, like a library or program or driver that's shared etc having issues because it's not fully shared, it's only partially shared or possibly 2 different versions with the same names etc.

Have you looked in the Event History? Checked for critical errors? Something not loading that should?

 

[guysbeingdudes]

Update to anyone who finds this thread:

From further troubleshooting, debugging, and research it seems like the root cause of the issue was a defective CPU. I have RMAed the Ryzen 3950x and haven't had any freezes occur since receiving the new CPU (with C-States enabled).