Windows 10: "Your PC encountered a problem and needs to restart." - Round 2.

WHEA Hardware Error again, but this time with an "Bus error":

Code: 0x00000124: WHEA_UNCORRECTABLE_ERROR (20.03.2017 07:22:49) [Windows 10] CAUSED BY: AuthenticAMD WHEA Notify: Machine Check Exception Type: BUS error Error: BUSLG_OBS_ERR_*_NOTIMEOUT_ERR (Proc 0 Bank 4) PROCESS: System Usual causes: Hardware, Incompatibility, May be driver corruption[/quote]


0x124



WHEA_UNCORRECTABLE_ERROR

A 0x124 is one of the worst STOP codes to encounter for the sole reason that the dump files usually give nothing away as to the cause of the problem. WHEA (Windows Hardware Error Architecture) errors signify a problem with hardware but very rarely pinpoint the culprit. In these scenarios it is advised to run a series of hardware stress and diagnostic tests to try and pinpoint the problem. A template is offered below which covers the four main components (GPU, CPU, RAM, HDD) and gives you a fighting chance of narrowing down the problematic device.

Generic "Stop 0x124" Troubleshooting Steps:

1) Ensure that none of the hardware components are overclocked. Overclocking means pushing the components beyond what they were designed for. If you do not know what that is you probably are not overclocking so go to the next step
2) Ensure that the machine is adequately cooled. If this is a laptop use compressed air to carefully blow out the heat pipe and fan when the computer is turned off. If a desktop take the side cover off and blow a fan on the components
3) Update all hardware-related drivers: video, sound, RAID (if any), etc. Do not rely on windows when it says the most recent driver is installed. It may be the most recent but it may also be corrupt.
4) Update the motherboard BIOS according to the manufacturer's instructions and clear the CMOS. Check with the computer maker for directions on this procedure and as usual it is always a good idea to back up your data.
5) Update ALL OS Windows updates
6) Stress test the major components. Start with RAM, CPU, HD, etc.
For RAM, use Memtest. Instructions for which can be found here.
Computers are extremely sensitive to problematic ram so any errors on memtest should be considered an issue and even a clear report for less than 8 passes can give a false negative.
For CPU use Prime95
For hard drives, use CHKDSK /F finds any problems on the drive(s), notably "bad sectors". You should also go to the HD maker and download and run their HD checking utility
7) Perform a "vanilla" (clean) re-installation of Windows: Install nothing that is not from the OS (not even Malware) until you have seen that the computer is not crashing in this state.
When the vanilla installation has run long enough that you are convinced it is OK, start installing updates and applications a few at a time, and wait until you are again convinced it is OK. If the crashes resume it is obviously the last group of installations that is the issue and remove them.
8) Re-seat all connectors, ram modules, etc. You can use the same can of compressed air to clean out the RAM DIMM sockets as much as possible.
Only attempt this if you are FULLY knowledgeable about the procedures.
9) If all else fails, start removing items of hardware one-by-one in the hope that the culprit is something non-essential which can be removed.






Diagnostic Test

RAM TEST


Run MemTest86+ - Test RAM - Windows 10 Forums



Note MemTest86+ needs to be run for at least 8 complete passes for conclusive results. Set it running before you go to bed and leave it overnight. We're looking for zero errors here. Even a single error will indicate RAM failure.

Make a photo of the result and post it.






Diagnostic Test

GPU TEST


Run FurMark - GPU Stress Test - Windows 10 Forums



Warning Your GPU temperatures will rise quickly while Furmark is running. Keep a keen eye on them and abort the test if overheating occurs.






Diagnostic Test

CPU TEST


Run Prime95 - Stress Test Your CPU - Windows 10 Forums



Warning Your CPU temperatures will rise quickly while under this stress test. Keep a keen eye on them and abort the test if overheating occurs.






Diagnostic Test

HDD TEST


Run HDTune to check health and scan for errors.



Note It may take some time, but look for signs of any errors or failure.

You can also run a disk check using chkdsk. Refer to the tutorial for details of how to do this.

 

[/quote] As I mentioned earlier in this thread, I have already run Furmark, CHKDSK, Prime95, etc, all showing no signs of a hardware problem. Since I posted my previous post here (but before you posted your most recent reply), I decided to uninstall Bitdefender, as I have heard claims that it can occasionally cause errors usually associated with hardware problems. I had originally wanted to put off uninstalling it, but have now decided that it's better to try to find out if the errors go away without Bitdefender before trying anything more drastic.

As usual, it's a waiting game, as there's no way to tell if a problem still exists unless it either reappears or ceases to cause trouble for so long that it's unlikely to still be around.

what is the particular significance of the error being labeled as a BUS error this time? Is there anything about it being a BUS error that could, in and of itself, lead to a clue as to the origin of the problem?

 

To understand what the error means, we need to decipher it a little.
BUSLG_OBS_ERR_*_NOTIMEOUT_ERR
LG stands for level generic, this is the cache position of the CPU where the error was noticed, as it is LG it means that the level of the cache couldn't be determined at the time of the crash.
OBS means that the error was observed as a 3rd party by the CPU core that noticed the error

What this means, the CPU picked up the error but wasn't involved in sending nor receiving the instructions.

 

I'm not very well-versed in this sort of highly technical information. Could you please explain the significance of that? Does it mean that the CPU itself is not to blame?

 

It basically means that the CPU isn't to be blamed, the CPU caught the error but wasn't doing much more than watching.

 

In such cases, what is the most likely culprit?

Could Bitdefender have possibly caused the crashes?

Could the fact that my RAM sticks are two different speeds (one 667 Mhz and one 800 Mhz), which causes my system to underclock the faster one to the speed of the lower one, have anything to do with it?

 

I can't say, I've only seen a single dump in the latest zip. With 0x124 there are multiple dumps required for patterns to be identified to possibly identify the culprit.

 

Here are all three from the three crashes that I recorded the data from:

Attachment 128381
Attachment 128380
Attachment 128379

Could you please try to find out if you can identify a pattern (and thus a possible culprit) from them?

 

How come there is only 1 dump in each zip? Do you cleanup your system with utilities like Ccleaner?


Reading your previous thread, you ran the hardware tests but I cannot find back certain relevant information such as temps.
Furmark isn't a test that shows errors when the GPU is unstable, it is meant to identify cooling issues and if the GPU appears unstable you should see artifacts or improper graphics on the screen.

For the mentioned reason, I would ask you to run the tests again and provide photos/screenshots or otherwise mentioned details about the test results.
In regards to the hard drive tests, I would ask you to run HDTune instead of Western Digital Data Lifeguard, Western Digital Data Lifeguard honestly sounds like a tool similar to SeaTools to me whereas HDTune provides more information in other areas.

I have removed Memtest86+ since you already ran it and there is only 1 result to get from it (errors or no errors).






STOP 0x124 Troubleshooting

Read carefully before proceeding.

*Arrow Run Prime95 - Stress Test Your CPU - Windows 10 Forums & CPU - Stress Test Using IntelBurnTest - Windows 7 Help Forums

*Arrow Run HDTune to

• check the health,
• scan for errors, no quick scan but full scan
• run a benchmark.

It may take some time, but please take the time you need to perform it properly.
When above is done please make screenshots of the following

• the health,
• the error scan,
• the benchmark incl. following
  • transfer rate,
  • access time,
  • burst rate,
  • cpu usage.

*Arrow Run SeaTools DOS to check the integrity of your HDD. SeaTools for DOS and Windows - How to Use - Windows 7 Help Forums
Run the long test.
*Arrow Run chkdsk
Disk Check - Windows 7 Help Forums
Use option TWO with parameter /r
Upload the chkdsk log *Arrow Check Disk (chkdsk) - Read Event Viewer Log - Windows 7 Help Forums






FOUR

GPU TEST
*Arrow Run FurMark - GPU Stress Test - Windows 10 Forums



Note Run Furmark for around 30 minutes.


Warning Your GPU temperatures will rise quickly while Furmark is running. Keep a keen eye on them and abort the test if overheating occurs.
Make a photo of the result and post it.

In Furmark you can find a tool called 'GPU-Z' (if you want to install GPU-Z is up to you), when GPU-Z is open click on the 'Sensors' tab, on this tab you can monitor

• the frequency of the GPU,
• the GPU memory clock,
• the temperature,
• the power,
• the load on the GPU,
• Etc.

I would like you to make a screenshot of GPU-Z after you have either cancelled Furmark for any reason or Furmark is finished, post this screenshot in your next reply with the result of Furmark.

 

No. I entirely restored my system to its pre-crash state using Macrium system image backups after each crash. I always saved the crash data (dump files, etc) before restoring my system, though. As I restored my system to its pre-crash state each time, I eliminated any possible operating system damage caused by the crashes themselves.

I know. It seemed to work fine when I ran it. I didn't see any artifacts. I didn't run GPU-Z, though. Are there any risks to installing GPU-Z?

Regarding temperatures, I eventually discovered that my CPU would often get dangerously hot (upper 60s and lower 70s Celcius) whenever the built-in AMD turbo core option (which automatically overclocks my 3.4 Ghz processor to 3.6 Ghz when it thinks I need the extra power) kicks on. As the turbo option can't be disabled directly on my system, I set the maximum CPU power to 99% in Windows Power Options, which keeps the turbo core from turning on, and thus eliminated my overheating problem (albeit with a slight sacrifice of CPU power, as it underclocks my CPU to 3.0 Ghz). I've also replaced the thermal paste, but that doesn't seem to have made much difference. I've read that high temperatures when the turbo core kicks on are very common, especially on computers with a stock cooler (such as mine).

Regarding the various other tests (CPU, hard drive, etc) it might be a while until I can run them again (I've been very busy with other matters recently), but I'll have to see when my schedule will allow them.

Should Prime95 be run with the processor power turned down to 99% in Windows Power Options to avoid overheating, or should it be run with the options set to 100% (which would almost definitely push it over 70 degrees at times)? The first (and only) time I ran Prime95, I had not yet changed the Windows Power Options setting down to 99%, so my CPU (which was running at full 100% capacity with the turbo core enabled) stayed hot during the test (lower 70s). That said, Prime95 found no errors even during that intense workout. I'm not sure how well running it with the maximum CPU power turned down to 99% (the way I have it now) would work. It would lessen any heating issues from the test, but would maybe make the test less effective. What would you recommend?


I doubt that it could be a hard drive issue (as I've already run CHKDSK and various other tests, plus my hard drive is almost brand new). Is running the same hard drive tests that I already ran (such as CHKDSK) really necessary, or would it be acceptable only to run the hard drive tests that I haven't run already (such as HDTune)?

 

The same risk as installing any other software you would normally use.

Your description about the CPU sounds like a cooling issue.

For chkdsk, simply provide the log.
AFAIK you only ran chkdsk, no other test has been already run.

 

It is, in a way. However, as I explained, the overheating only happens when the AMD turbo core turns on, and I've read that this is a fairly common (and apparently normal, albeit problematic) issue with AMD turbo cores, especially in computers with stock coolers. You see, the turbo core is a feature that causes the CPU to use an extra 0.2 Ghz when it thinks it needs to. When the turbo core is disabled (something that can be done in the BIOS of some computers, but apparently not in my computer's sparse and basic BIOS), the overheating usually goes away.

I got around this issue by setting the maximum CPU power setting in Windows Power Options to 99%, which causes my CPU to only use up 3.0 Ghz, completely eliminating the heating issue. My CPU temperatures are now usually in the 50s Celcius, sometimes in the 40s when using less programs, and ocassionally even in the upper 30s when idling.


You didn't tell me whether I should run Prime95 with the Windows maximum CPU power setting set to 99% (as I do now, which eliminates heating issues), or at 100% (which is where it was when I first ran Prime95, but can and did cause high temperatures, but might provide a more thorough test). Which would you recommend?

Here it is:

Attachment 129492

I've also run the long test of Western Digital Data Lifeguard. It's basically the Western Digital equivalent of Seagate's SeaTools, and is supposed to provide more accurate results than SeaTools on Western Digital drives (such as my C drive). And I also ran all of the following Windows commands (none of which found any problems):

dism /Online /NoRestart /Cleanup-Image /CheckHealth

dism /Online /NoRestart /Cleanup-Image /ScanHealth

dism /Online /NoRestart /Cleanup-Image /RestoreHealth

sfc /scannow

Is running SeaTools necessary at this point (as I've already run Data Lifeguard)?

 

I would suggest to run at 99%

Yes, it's similar but not the same. (Sometimes I'm a real jerk with this kind of things...., for which I apologize)
Reading more about it you can skip the SeaTools test.

 

I've now run the tests.

I ran Prime95 for over 5 hours. There were no problems. Here are two screenshots (one from while Prime95 was still running, and one from after I stopped it).

Attachment 130104

Attachment 130105

Here are the long error scan scan results from HD Tune Pro:

Attachment 130106

Here is the drive health information (which I believe to be HD Tune's reading of the drive's S.M.A.R.T. data):

Attachment 130117

Here are the results of the benchmark:

Attachment 130108

As an added bonus, here are the S.M.A.R.T. results from CrystalDiskInfo and Western Digital Data Lifeguard:

Attachment 130121

Attachment 130124

Furmark seemed to me to run fine. Here is a screenshot of Furmark made after it had been running for 30 minutes:

Attachment 130109

Here are two screenshots of GPU-Z made while Furmark was still running (after making the screenshot of Furmark):

Attachment 130110

Attachment 130112

Here are two screenshots of GPU-Z made after I closed Furmark:

Attachment 130114

Attachment 130115

 

Please fill in your system specsPlease follow this tutorial and download the tool. The tool will give you detailed information about your system specs, please fill in your system specs more completely including PSU, cooling and other used stuff like mouse, keyboard, monitor, case, etc.
The PSU, cooling and other stuff are NOT mentioned in the tool.
In the left corner below in your post you find 'My System Specs'. After clicking it you can find a link a little below that says 'Update your System Spec', click on this link to get to the page where you can fill in your system specs.