Lockup during memtest

[Cipher]

Hi,

My PC has been freezing during some games, no bluescreens, just a horrifying noise from the speakers and a completely unresponsive PC. After ruling out my GPU, I suspected my RAM.

After approximately 6 hours with 0 errors, the system locked up. I turned off my monitor during testing, but found that when I turned it back on it wasn't getting a signal. At first, I thought for some reason memtest just doesn't allow my monitor to regain signal, but I've since run it again and successfully turned it off and on.

So, presumably, my RAM is indeed faulty? Any advice would be great.

Thanks,

- Elliot.

[Cipher]

I've since looked at the logs, and it seems like the test was still running at the time I manually switched off. For some reason my monitor just refused to reconnect to the system. I did try the key commands to end the test, but they also didn't seem to do anything, hence why I suspected a freeze.

Really scratching my head now.

[David (PassMark)]

There is no sleep / power save function in MemTest86. The monitor should not turn off by itself. So maybe just run the test again with the monitor on.
Don't know why the monitor didn't recover.


If MemTest86 doesn't show any errors. Try,
- Checking for overheating while gaming & cleaning out any built up dust
- Check video card drivers are up to date
- Check the PSU. But it is hard to check for intermittent problems without swapping the unit.
- If the problem only happens in one or two games, it might be a bug in the game itself.

[Cipher]

Hi David, and thanks for your prompt response.

Drivers are all up to date, with previous video drivers removed using DDU. As for temperatures, my CPU rarely strays above 45°C, and my GPU appears to peak at around 70-72°C. (It's a reference blower model, so not ideal for cooling)

My PSU has been a concern of mine. It's a Corsair RM1000i that's less than a year old, which makes me doubt it has a problem, however.

Funnily enough, I have only experienced the crash in 2 games, namely Overwatch and the Witcher 3. The fact that the crash is identical in both is what made me first assume a driver, and later, a hardware problem. It is also incredibly rare, sometimes only occurring after several continuous hours of usage.

I'll probably run each stick individually overnight through memtest86, just in an attempt to rule them out for sure. What's frustrating is it could be almost anything at this point.

Apologies for rambling so much about non-memtest related topics - I'm just at my wits' end trying to hunt down a solution.

[Lilien]

Hello,

I turned off my monitor during testing, but found that when I turned it back on it wasn't getting a signal.

I've seen the same problem with my new build (Gigabyte Z170 Gaming 7 with I7-6700K and 16GB Kingston HyperX Memory) once with an older 4.3.x build (UltimateBootCD) and now with the recent v7 beta.
I turned the monitor off before I went to bed and let MemTest86 do the job overnight and in the morning the monitor had no signal.
The run from the UltimateBootCD gave me no logfile but I have the logfile from the second run with the v7 beta, these are the last entries:


2016-07-14 07:46:46 - Cleanup - Unlocking all memory ranges...
2016-07-14 07:46:46 - MtSupportRunAllTests - Test execution time: 2827.722 (Test 13 cumulative error count: 236)
2016-07-14 07:46:47 - Finished pass #4 (of 4) (Cumulative error count: 566)


Beside the the errors, this seems to be a normal end, but what is going on to cut the video (integrated graphics)?

Regards Juergen

[keith]

That does sound a bit unusual. The fact that it happened on versions of MemTest86 that run on different platforms (UEFI, BIOS) makes me suspect that it could be a BIOS issue (The UEFI version of MemTest86 was almost completely rewritten from the legacy BIOS version). In either case, no power/video management is done during testing.

In the case of MemTest86 v7 beta, there should be a prompt indicating that the test has completed and it should be indefinitely waiting for the user to press a key to continue.

[Lilien]

That does sound a bit unusual. The fact that it happened on versions of MemTest86 that run on different platforms (UEFI, BIOS) makes me suspect that it could be a BIOS issue

I tested this further and you are absolutely right.
The DisplayPort signal is gone in the moment the monitor is shut off and it never comes back after turn on ("No signal"). Its the same behavior in the UEFI Bios, so the problem seems to be based on UEFI.

Regards Juergen