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

[David (PassMark)]

Can you also try the 7.0.1003.zip release Keith linked to above.

In the V7.0 release of MemTest86 we changed the internal clock we are using to measure various things. As background: a modern PC has a variety of internal timers and clocks that can be used by software. Any misbehaviour by the clock can result in things happening at the wrong speed or measurements like the CPU clock speed being wrong. For a long long time MemTest86 used a timer in the Programmable Interval Timer (PIT) chip. Every PC made in the last 20+ years has one of these and they all had the same interface, until recently. A very small number of new PCs don't have a PIT. So in MemTest86 V7 we now use the High Precision Event Timer (HPET) by default, and the PIT if a HPET can't be found.

Unfortunately on a few machines, the HPET was available didn't return values we expected, which we suspect caused some of the above issues. In the next patch release we are a bit more careful in using the HPET, plus we added fallback mechanism to use the legacy PIT to measure the clock speed if the measured CPU clock speed using HPET is unreasonable.

[zerophase]

Happens on 5960x, as well. The issue continues with 7.1. Getting stuck at Testing multiproccesor support for at least a good five minutes before deciding to just restart. I even tried with compatibility support turned on for bios boot, and that didn't help either.

[keith]

zerophase, do you have the logs for your MemTest86 run? It seems likely to be a UEFI BIOS multiprocessor issue similar to the ones listed here:
http://www.passmark.com/forum/memtes...election-modes

If you can upload or e-mail a copy of the Memtest86.log file under the EFI\BOOT\ directory of the USB drive we can confirm if that's the case.

[Psychor]

I created a bootable USB of MemTest86 V7.1 Free Edition today. When I booted it via UEFI on a new 6900K machine I just built I had a similar issue - it just sits at "Testing multiprocessor support..." and doesn't budge. I didn't have the patience to wait 20 minutes like the above poster though. I checked the USB stick and there were no logs in any of the folders.

UPDATE: I just took the same exact USB stick and put it in my 3930K system, booted with UEFI and had no issues. Wonder why it's having an issue with the newer processor.

UPDATE2: I ran diskpart, cleaned the USB and then re-imaged MEMTEST86 with Rufus. I can now get passed the "Testing multiprocessor support..."

[zerophase]

zerophase, do you have the logs for your MemTest86 run? It seems likely to be a UEFI BIOS multiprocessor issue similar to the ones listed here:
http://www.passmark.com/forum/memtes...election-modes

If you can upload or e-mail a copy of the Memtest86.log file under the EFI\BOOT\ directory of the USB drive we can confirm if that's the case.

My motherboard, Rampage V Extreme, isn't listed in that list.
Here's the paste(It includes logs going back a couple versions): http://pastebin.com/r9DVexGD

[Rodsman]

So I built my first system last May and have had random bsod and the error code indicated it might be my ram. I bought the pro version of the test and ran it yesterday. It showed over 30000 errors which seemed crazy to me. Rerunning it today and it has shown no errors but it's been going for 8 hours? Clearly I'm new to all of this and have the same CPU in this thread. How long should all the tests take and any thoughts on the huge difference in results. Thanks. I know those of you that know way more than I do may want more info but I wanted to see about the timing issue before wasting anyone's time.

[keith]

Without a screenshot or log, it is hard to say what is causing the errors. It may be an intermittent error due to external factors or it could be a BIOS error.

The time it takes to run the MemTest86 depends on several factors. On average, it may take about 1-1.5 hours to complete the first pass. See this page for details:
http://memtest86.com/technical.htm#execution