Interesting issue running Memtest86 v7.0 Free with i7-6700K

[WtheFourth]

So, I'm trying to figure out what the heck is wrong with my new rig. I get these "stutters" so to speak, when all audio coming from and being sent to my computer is incredibly delayed (think a solid half second between each sound), my mouse seems to be moving dramatically slowly (skipping numerous pixels when sweeping across the screen), any moving image on the screen, from games to loading bars to whatever, becomes incredibly choppy, etc. It's not easily able to be reproduced, as I can't figure out what triggers it. Until...

I made a bootable USB of the newest version of Memtest86. I booted to it and suddenly, my computer stalled horribly when it ran the initial system info checks prior to booting into Memtest86. Like, it was stuck on "Measuring CPU/cache/mem speeds..." for about 15 minutes, and "Testing multiprocessor support..." for another 20 to 30 minutes. So, that was the first instance in which I thought something might be wrong with either Memtest86 or my CPU or my motherboard or... something.

So, it finally booted into Memtest86 after over 30 minutes, and my Core i7-6700K clock speed was 400.0 MHz. Boost clock was 450.0 MHz. So, it was off of what it should be by about 10 times, which would explain the slowness. I figured it may still be a Memtest86 issue, so I went ahead and ran the test, until it got to the sleep test and took several minutes to tick from 300 seconds to 299 seconds. So, I'm trying to figure out what the heck could be going on. I restarted my computer after that, and booted into Windows 10 and ran the Windows Memory Diagnostic, which, naturally, found no errors.

Now I'm at an impasse. I know that something is wrong, but I don't know what. I think it is probably either my RAM or my motherboard, but I can't be sure which.

Parts:

CPU: Intel Core i7-6700K
RAM: Corsair Vengeance LPX DDR4-3000
Motherboard: ASUS Maximus VIII Gene
Storage: Samsung 950 Pro M.2 SSD
GPU: EVGA Geforce GTX 1080
CPU Cooler: Corsair H100i V2
PSU: EVGA 850 G2

[David (PassMark)]

Maybe CPU is massively overheated and throttling its clock rate down to avoid cooking?
e.g. heat sink on CPU has fallen off.

[WtheFourth]

It doesn't seem to have overheated, as the board is registering temperatures around ~24-28°C on idle and ~50-60°C on full load. The highest I've ever gotten the CPU was ~78°C during a stress test. After the "stutter" happens in Windows, I typically check HWMonitor, which shows my maximum temperature still being well below anything dangerous.

I've noticed the "stuttering" happens typically when loading applications, but I'm not sure what the root cause would be.

[bcs]

Disconnect the M.2 SSD, take out the GTX1080 (and all other PCIe cards that you might have connected), remove all but 1 DIMM, and Reset BIOS to defaults. Then run Memtest again and see if the "Measuring CPU/cache/mem speeds..." and "Testing multiprocessor support" still takes 20 to 30 minutes. If not, then add back more RAM and retest. Then add the M.2 SSD and retest, then the GTX1080 and retest.

[keith]

It would be good to see the logs from the MemTest86 run if you still have them.

[WtheFourth]

Unfortunately, I don't know that any of the logs saved given that I canceled the test when it took several minutes to count down from 300 seconds to 299 seconds on the Sleep test.

The plan for either tonight or tomorrow, depending on when I have time, is as follows, in order:

* Reseating CPU
* Reapplying thermal paste (I use Thermal Grizzly Kryonaut)
* Removing GTX 1080
* Removing M.2 SSD
* Removing one stick of RAM
* Clearing CMOS
* Testing Memtest86 to see if it boots properly
* Re-adding second stick of RAM
* Rerunning Memtest86 to see if it boots properly
* Potentially running running full test to verify RAM is fine?
* Inserting M.2 SSD to see if Memtest86 boots properly
* Inserting GTX 1080 to see if Memtest86 boots properly

I think that should at least isolate the issue... provided the issue isn't the motherboard itself. I've got RMA paperwork handy, just in case.

[keith]

Unfortunately, I don't know that any of the logs saved given that I canceled the test when it took several minutes to count down from 300 seconds to 299 seconds on the Sleep test.

MemTest86 starts collecting logs automatically at the moment it boots up and is saved as MemTest86.log under the EFI\BOOT\ directory of the USB drive. Anyways, if you have it, it would be good to have a look and see if there is anything useful in there.

[WtheFourth]

MemTest86 starts collecting logs automatically at the moment it boots up and is saved as MemTest86.log under the EFI\BOOT\ directory of the USB drive. Anyways, if you have it, it would be good to have a look and see if there is anything useful in there.

Oh, awesome, I definitely still have the logs then and I will upload them this evening!

[WtheFourth]

Update: I've done everything above, up to and including resetting the CMOS battery - still got the same issue, leaning heavily on it being something with either the CPU socket or the CPU itself. It seems unlikely it'd be the CPU, and I typically suspect ASUS before ever suspecting Intel. I'm currently only running the CPU, CPU cooler, and a single stick of RAM. On first boot, CPU was at 27°C, still stuck on testing multiprocessor support. Radiator is showing that it was running at 1500 rpm, give or take, in BIOS menu. CPU fans are pumping air as expected. No USB devices aside from keyboard connected.

So, motherboard?

Sidenote: will grab logs using laptop ASAP.

[WtheFourth]

Log: http://pastebin.com/W9dDXLEm

Video I recorded to show some of what is going on in Memtest: https://youtu.be/E76TfwBjbQo

Here's an image showing CPU operation within Windows (the high of 53-56C was while running the stress test in the Intel Processer Diagnostic): http://i.imgur.com/uM2eSpp.png

[keith]

Thanks for the uploads.

Can you give the following build a try:
http://www.passmark.com/ftp/memtest86-usb-7.0.1003.zip

Also, if you have another fast, reliable USB memory stick, can you try burning the MemTest86 image on that instead. We have seen cases where an older USB drive caused MemTest86 to slow down considerably due to the amount of time it takes to write entries into the log file.

[WtheFourth]

Awesome, thanks. I bought and used the memory stick I'm using now to flash the BIOS on my motherboard on around 5/13/2016, but I will try both this one and a new one after I run to Best Buy this afternoon.

The existing one is a Sandisk Cruzer Fit 16GB, and I have a Silicon Power J80 16GB (although it currently is a Windows 10 Install drive - I can probably overwrite it for now though). Any specific brand or anything you'd recommend?

[WtheFourth]

I'm unable to boot to the UEFI version of Memtest linked above, but I am running v4.3.7 successfully as I type this. So, progress?

http://imgur.com/b1xsXzZ

Edit: turned off secure boot, successfully running UEFI version. Hopefully I can find some errors so I can figure this out!

[WtheFourth]

Good news and bad - Memtest86 successfully completed 10 passes on both sticks of RAM individually, so that means the memory is alright! However... that means something else is likely the problem (welp), should the issue persist when I hook everything back up and try out Windows. I suppose I'll leave HWMonitor running to see if I have any outrageous dips or anything.

[tsulkoske]

Wanted to chime in because I'm having the same exact problem with the same exact processor. It would get stuck on test 10 and basically freeze during the countdown. Version 4.3.7 works fine though! Several passes in and no errors yet.