Does memtest 7.4 shut the computer down at all for any reason?

[mikeymikec]

I haven't used it a great deal (I used memtest86+ 5.01 for a long time), so I can't be 100% sure what to expect from it.

 

[dave_the_nerd]

I don't recall it doing that. I recall computers freaking out and powering off because of heat or bad RAM.

Watch me be wrong though.

 

[mikeymikec]

I don't recall it doing that. I recall computers freaking out and powering off because of heat or bad RAM.

Watch me be wrong though.

Yeah, I strongly suspect it doesn't do that by design. I think I'm dealing with another flaky CX430 PSU, though it's holding up quite well with memtest86+ 4.2 (god knows where my 5.01 CD is)... maybe it's because 4.2 only uses one CPU core so therefore the machine is less loaded?

I have seen memtest86 be problematic on certain configurations before though (maybe just when newer hardware comes out and an older version disagrees with it), though usually in that situation it reliably hung after x seconds, whereas on this PC it's all over the place.

 

[Dankk]

Just like the older FOSS version of Memtest86+, PassMark's own version of Memtest86 (which is actually still maintained and updated) should behave similarly, in that it should just run forever until you shut it off. Edit: I lied, it looks like Memtest86 only runs a few passes by default, and then finishes.

It's true that Passmark's Memtest86 contains a lot more features than the older one (such as multi-core support) so there may be something in there that is causing your machine to break. I would doubt that it's actually the fault of Memtest86 though; as you said, there may already be something wrong with your machine, and Memtest is simply the catalyst for making it break.

 

[mikeymikec]

The computer handles lots of passes with memtest86+ 5.01 with SMP enabled no problem, but regularly barfs (screen either goes black or freezes) with memtest 4.3.7 or 7.4 (it hasn't got past 2 hours with any version of memtest non plus). It's not the PSU, same thing happens with a spare. I'm wondering whether there's something odd about this board and memtest which isn't necessarily a hardware fault.

 

[DaveSimmons]

Could the CPU be overheating?

 

[mikeymikec]

Could the CPU be overheating?

Nope. About 40C while memtest was running.

 

[Shmee]

What are your specs? In my experience, the UEFI memtest can have bugs/freezing on certain platforms.

 

[mikeymikec]

i3-3225
ASUS P8Z77-M PRO
4GB DDR3-1333

V4.3.7 is the version intended for non-UEFI platforms (7.4 for UEFI).

 

[Shmee]

Hmm, it should be fine with that system, as far as bugs and possible incompatibilities. Typically the issues tend to arise on X99 and newer HEDT systems, certain server systems, and Ryzen/TR systems.

I assume you did not find any errors though? Like you suspected, it could be PSU related, or possibly board or CPU related.

 

[mikeymikec]

No errors found. Except in a dream I had the other night: The familiar red line popped up on the memtest screen, I thought "ooh!", then immediately woke up and thought, "crap.".

Prime95 blend for a couple of hours: OK
Memtest86+ for about 5 passes (4 or 5 hours IIRC): OK
Windows memory diagnostic: OK
Firefox FishIE tank for about an hour (I checked that it was giving the iGPU a good workout): OK
full chkdsk on SSD: OK

The original symptom was that the screen would go blank or freeze during use. I found that even when sitting idle it would do it too. The original symptom could easily have been graphics related so I updated the driver. As the system was upgraded through the free offer from Win8x to Win10, I took a look at the storage driver which was dated 2013, and as no Win10 RST driver is apparently available for it I switched over to the standard storahci one. Since I did those things it isn't been problematic in Windows, but I guess I wasn't overly convinced that a Win10 system that has had no recent major updates (KB4462919 was installed on the date that it started having problems though, mental note to look into compatibility issues with that) installed recently should out of the blue start acting weird so I suspected hardware. I wonder whether I've been barking up the wrong tree.

Since those driver updates I've mostly been letting Windows run idle for several hours at a time and disabled sleep mode, no problems.

I hate intermittent issues.