Computer freezing and restarting

[MK9]

Hi,

I'm having a problem since yesterday with my desktop PC. Basically after playing a game for more than 30-45 mn the game would freeze and windows would become unresponsive. Pressing the windows key did nothing and cntrl-alt-del had no effect either.
I tried a different game and the same thing would happen.

After the freeze the PC gets a brief BSOD and reboots. I then get a message telling me to choose a boot device.

I just ran Furmark to check my temperatures and everything was normal, but the PC froze and crashed immediately afterwards. I noticed that as it crashed the HDD light stayed on all the time.

Before this I ran a memtest86 and it found nothing. I also tried WhoCrashed but there are no dumpfiles available.

CrystalDisk tells me the HDD and SSD are in good health. I don't know what the other details mean but I can take a screenshot.

Here's the specs:

MB: Asus p8-Z68V GEN3
CPU: i5-2500k (OC to 4.3Ghz using the MB setting)
RAM: 2 sticks of 4GB DDR3-1600 Corsair
GPU: ASUS Radeon 7970 (OC to Ghz levels)
SSD: Crucial M4 SSD 64GB
HDD: Seagate Barracuda 1TB
PSU: Corsair AX750

Any tips would be appreciated

 

[Smoove910]

sounds like a possible unstable overclock... have you tried running the same scenario without the overclock to see what happens?

In the AMD world, I would typically see if adding a sliver of voltage would stabilize it.

 

[MK9]

sounds like a possible unstable overclock... have you tried running the same scenario without the overclock to see what happens?

In the AMD world, I would typically see if adding a sliver of voltage would stabilize it.

Thanks, I just tried it but the problem persists:

I just reset all the OC values to default, ran a furmark and then launched a game. About 40 minutes in there was a slight freeze that lasted a few seconds. Then the game resumed. I alt-tabbed to check something else and got stuck on a black screen with the HDD light fully on.

 

[chusteczka]

We need to establish a foundation:

1. ensure stable power supplied,
2. check and install latest bios,
3. test memory,
4. check and install latest chipset and video drivers,
5. system test and monitor temperatures,
6. test cpu overclock,
7. test video overclock.

BlueScreen Restart
Turn off the Automatic Restart upon a system crash.
If the system bluescreen crashes again, what does the error message point to?

Electrical Power

1. How old is the electrical wiring in your home or building?
2. Do you have an Automatic Voltage Regulator (AVR) or battery backup? If yes to either or both, what is the wattage capacity and what all is plugged into it?

Bios

1. Do you have the most recent bios installed? Asus P8-Z68V GEN3 Support

Testing Memory

1. Set bios OC settings to default for both processor, memory, and video.
2. Run Memtest through a single pass for each individual memory stick. Then follow up with two full passes or overnight with all memory sticks. In my experience, I prefer Memtest86+. Memtest86, Memtest86+

Chipset and Video Drivers

1. Ensure the latest motherboard chipset drivers are installed. Asus P8-Z68V GEN3 Support
2. Ensure the latest video drivers are installed. AMD Video Drivers

System Test

1. Run Prime95 a minimum of 6 hours, preferably 10 or 12. What happens? I am expecting either a calculation error or a whole system crash at this point.
2. If no errors, test with OCCT for a few hours. (I have not used this program in a few years. Not sure what to suggest with it.)
3. What are the motherboard, CPU, and video temps as provided by CPUID HWMonitor? Make note of any temperature above 60C.
4. At this point, it would be good to use your system normally for a week or so to see if any crashes continue to occur. Continue monitoring temps for any temp above 60C. May run Prime95 overnight to further ensure stability.

Testing CPU Overclock
If memory has tested fine and the system is running well, then you may test a CPU overclock. Ensure to set video to default bios settings, no overclock on video.

Testing Video Card Overclock
If the system is running fine with no issues, then a video overclock may be attempted.

 

[Smoove910]

what voltage does your ram require? 1.5 or 1.65v?

 

[Burpo]

Sounds like a weak power supply..

 

[MK9]

what voltage does your ram require? 1.5 or 1.65v?

It's 1.50v

Thanks chuzstecka, I'll try those tests tomorrow.

Sounds like a weak power supply..

I've had the PC since 2011 and only changed the GPU a bit more than 6 months ago, and I never had a problem. 750W was always enough for it.

Is there are way to check how much watts the PSU is generating?

 

[Ketchup]

I would be interested in hearing about how it responds with CPU at stock, video card at stock, and the memory at either SPD or one click lower (1333). If it fails at that, I would lean towards Burpo and a possible power supply issue.

 

[chusteczka]

A 750w PSU should be enough for that system but it is possible the PSU is failing in some way. It would help if MK9 could borrow or replace the PSU for testing or purchase and use a PSU tester. That is the primary reason I am wondering about the power being supplied to the computer.

newegg.com - Power Supply Tester

 

[MK9]

Tried again with stock speeds and memory frequency at 1333mhz, same result.

Will try to find a PSU meter, thanks.

 

[chusteczka]

A PSU meter and an Automatic Voltage Regulator (AVR).

OPTI-UPS SS1200 Voltage Regulator ($32)
OPTI-UPS SS2000 6 Outlets Home AVR Series Automatic Voltage Regulator ($46)

The Opti-UPS S1200 = 600 watts.
The Opti-UPS S2200 = 1000 watts.

Here is the manufacturer product specification sheet.
http://www.opti-ups.com.tw:8080/User manual/0401SS1200~2000 User Manual.pdf

 

[Loser Gamer]

When I got my video card it worked for a few months then it started to give me memory dump crashes as I started to play games. I solved it by underclocking -13 on GPU core and memory clock. I used MSI Afterburner.

I'd be very interested if you tried this to see if it fixes your issue since they read similar. Turn up fan speed also some on GPU with Afterburner.

 

[MK9]

A PSU meter and an Automatic Voltage Regulator (AVR).

OPTI-UPS SS1200 Voltage Regulator ($32)
OPTI-UPS SS2000 6 Outlets Home AVR Series Automatic Voltage Regulator ($46)

The Opti-UPS S1200 = 600 watts.
The Opti-UPS S2200 = 1000 watts.

Here is the manufacturer product specification sheet.
http://www.opti-ups.com.tw:8080/User manual/0401SS1200~2000 User Manual.pdf

I can't seem to find them here (I live in Ireland)

My PC is connected to a surge protector but no AVR.

When I got my video card it worked for a few months then it started to give me memory dump crashes as I started to play games. I solved it by underclocking -13 on GPU core and memory clock. I used MSI Afterburner.

I'd be very interested if you tried this to see if it fixes your issue since they read similar. Turn up fan speed also some on GPU with Afterburner.

I actually went ahead and re-installed my old GTX 570 with standard clocks and without any CPU OC either.

After 45mn I got the same crash, this time a blue screen and I was able to write the codes down:

STOP: 0x000000F4 (0x0000000000000003 , 0xFFFFFA8009EB4060 , 0xFFFFFA8009EB4340 , 0xFFFFF800037DE270)

Physical memory dump FAILED with status 0xC0000010


Thanks again for all your replies.

 

[chusteczka]

The AVR is helpful but not necessary, possibly a luxury. A good PSU with Active PFC correction will suffice. But if you have old electrical lines in your building (1970's or older) or frequent storms where the power drops for a moment and the clocks reset, then an AVR can certainly be useful. Often Uninterruptible Power Supplies (UPS with batteries) have AVR, but not always. But again, this is an extra item not necessary but can certainly be helpful.

The above STOP codes are memory addresses and not helpful (FYI) but the line after that is typically the verbose description for failure and can be helpful. In this case, it tells me of a hardware failure where the system dropped before the software could exit properly. This further tells me it was not a driver failure.

There is this other similar thread in which a slightly different approach was taken. Maybe reading through it can give you some ideas.
PC Randomly Freezes.

Could you provide two screenshots of CPUID HWMonitor; one at idle and the second at load for 7-10 minutes? Prime95 or OCCT can be used to place the system under load. My preference would be for Prime95, and select the option to heavily test the memory.

It might help to bump the memory voltage +0.05 volts, ie. from 1.50v to 1.55v.

You could also bump the voltage to the video card, or manually underclock it as Loser Gamer mentioned.

If the system was working and now it is not, then some piece of hardware is slipping, maybe not outright failing but on its way. The first item typically looked at is the PSU, and is relatively easily replaceable.

 

[MK9]

Idle:

After Prime95:

 

[chusteczka]

ALT-PrintScreen will provide a screenshot of just the application in focus.

Unfortunately, the Voltages section does not provide PSU voltages. That would have helped. However, of the other values provided, everything looks normal.

Have you tried bumping voltage to the memory and/or video? There are three possible components in my view; PSU, memory, and video card.

Maybe the southbridge chipset with hard drive access but your MB temperatures are fine... The southbridge could be tested by a manual virus scan while leaving HWMonitor open. See if the computer freezes then.

Another possibility is to open several applications like word proccessor, browser (Just not on an Adobe flash page, if it crashes you want to know it is not due to flash player.), game, and Prime95. Then switch between applications to test the processor's ability to access data related to the context switching.

It might help to obtain a new PSU.

 

[MK9]

I launched HWMonitor and a full scan by Windows Defender (I previously uninstalled a better AV because I thought it was might be behind the crashes) and when I came back the PC was frozen. I'll try that again tomorrow along with the other test.

Do I have to remove the MB completely to change the paste?

 

[chusteczka]

Possible Causes:

• software glitch - recent OS or software updates, recently installed programs, recently failed updates.
• drivers - chipset, video, ...
• motherboard voltages - bump memory voltage, processor, or NorthBridge to help stabilize.
• PSU voltages - Check with HWMonitor, possible failing power supply?
• temperatures - cpu, northbridge, southbridge; check with HWMonitor. Anything overheating?

You mentioned the system froze while performing a virus scan. That could mean the southbridge chipset is (possibly) overheating, due to continuous hard drive data access. But it does not rule out an issue with windows (OS files, drivers, programs). Yes, you would need to pull out the motherboard so you could squeeze the pins on the backside of the motherboard with needlenose pliers. Although we do not have indication of anything overheating. It might help to check the temps in the bios.

Did you try slightly bumping the voltages for cpu, memory, video, or northbridge? (ie. Memory from 1.50v to 1.55v)

Could you first download the Avira Rescue CD, copy it to a CD or USB drive, and scan your system with it? I believe this also has Memtest on it, could you then run Memtest (again)? If your system does not freeze or crash during these tests, then we know everything is physically fine for standard operation and the NB and SB chipsets are not overheating.
http://www.avira.com/en/download/product/avira-rescue-system

If the system does not crash during Avira or Memtest, then the problem is most likely with Windows (drivers, OS, or programs). I might help to bump the voltages for cpu, memory, video, or northbridge.

 

[MK9]

No crashes or freezes occured during the Avira scan or memtest86+.

Tomorrow I'll try bumping the voltages.

Thanks.

 

[chusteczka]

A non-Windows environment (Avira and Memtest86+) works fine. Continuous data flow through the southbridge chipset for the virus scan worked fine with no crash. This means the issue is most likely not chipset temperatures and there is no need to pull either chipset heatsink to replace the thermal paste.

The system crashes when pushed. You have not mentioned that the system has crashed when idle, only during prolonged gaming.

This leads primarily to voltages, secondarily to the PSU or possibly (?) system drivers.

Hopefully, bumping the voltages a little will help.

 

[MK9]

The system crashes when pushed. You have not mentioned that the system has crashed when idle, only during prolonged gaming.

Update: I had the same crash while at idle. I had a download running and had the music player on, that's all.

First the screen freezes (but the cursor still moves) then I get a sound loop followed by blue screen as before.

 

[Ketchup]

Sorry I have been away for a while MK9. Busy couple weeks at work. The error you mentioned in post #13 indicates a hard drive error.

You could do a long test using Seatools for DOS (they make a Windows version but I don't trust it's long test from within the OS) and see if any errors are correctable.

 

[chusteczka]

STOP: 0x000000F4 (0x0000000000000003 , 0xFFFFFA8009EB4060 , 0xFFFFFA8009EB4340 , 0xFFFFF800037DE270)

Physical memory dump FAILED with status 0xC0000010

Going off what Ketchup mentioned, I searched and found Blue Screen Bug Check Codes.

STOP: 0x000000F4
Interpreting a Bug Check Code
Bug Check Code Reference
Bug Check 0xF4: CRITICAL_OBJECT_TERMINATION

The CRITICAL_OBJECT_TERMINATION bug check has a value of 0x000000F4. This indicates that a process or thread crucial to system operation has unexpectedly exited or been terminated.

However, I am not sure of the exact meaning behind the hex values in parentheses and the "Physical memory dump FAILED ..." statement. Though, a google search does tend to point to hard drive failure, as Ketchup mentioned.

 

[MK9]

Seagate for DOS can't detect any hard drives

Here's the values from CrystalDisk if they are any help (I don't know what they mean)

 

[Ketchup]

Hmm, I was looking more closely at your board, and are you using the Jmicron controller for any of your drives? It would be very unusual for Seatools not to detect a mechanical hard drive, at least.

I would recommend you run all your drives on the Intel controller.