GSkill RAM RMA, new kit throws errors

[yenic]

I'm not crying yet because I haven't even reseated them, but the odds of getting back bad RAM has to be very, very low. I first set XMP II, ran all tests in memtest86+ and got errors. Then I set XMP I and did another all nighter, same thing.

 

Any recommendations on things to do? My CPU isn't overclocked, just the normal Intel boosts are enabled. Also got Intel standard power limits enforced, no MCE etc.

What are the chances the memory controller on my CPU is bad, or motherboard? 

 

Notice on both tests, it was core 8 on both tests that detected the error. I'm wondering if it's not my CPU. Maybe I should try to detune it and try again?

 

XMP II, the PC was frozen on this screen. I had to do a hard power off.

 

XMP I

[leclod]

You could loosen the Ram slightly, 3000MHz instead of 3200 or 15-15-15-35 instead of 14-14-14-34.

I'm running my 3600MHz Ram at 3500.

[yenic]

12 hours ago, leclod said:

You could loosen the Ram slightly, 3000MHz instead of 3200 or 15-15-15-35 instead of 14-14-14-34.

I'm running my 3600MHz Ram at 3500.

 

12 hours ago, wONKEyeYEs said:

Run memtest on RAM in another PC.

Thank you gents. I may already be out of the woods. My plan was 1) set BIOS optimized defaults and test, then 2) reseat RAM, then 3) cry about it.

So first step before reseating my RAM is I reset my BIOS to optimized defaults. Then I made only the bare minimum changes. I didn't put into place the entire list of changes that I usually do (which include CPU related settings). Just enabled Wake On LAN, set my fan speeds, disabled motherboard lights, and enabled XMP II. Touched nothing else. That should do the job in isolating the RAM setting change, and it should work under both the XMP 1 or 2 profiles.

Then ran memtest86+ all night with all tests enabled.
 

 

Over 12 hours no errors. So this leads me to believe my issue isn't the RAM. It was likely some other BIOS setting causing the CPU/memory controller to not play nicely in some situation that memtest86+ puts it in.

 

 

Basically my lesson learned, which I already did for many years now, is to run stability tests after changing any settings. People change BIOS settings, play games or whatever, and go on their way, and never realize that the options available in these systems give you plenty of power to hang yourself with them.

I've been guilty of this too often myself. But my next BIOS change will accompany an all-night memtest session for sure.

[yenic]

Another 11 hour test session with no errors. This time I made a CPU setting change, from MCE enabled to disabled (stock Intel power limits) and ABT (Adaptive Boost) enabled as I prefer.

 

The remaining settings I changed in the past but am not now, leaving on default/auto or whatever optimized defaults are. The issue may has somehow been related to something below. No need to set them as I have everything I care about already being set now, so I'm done with changes. But it must be something related to these. I am in Gear 1 mode anyway, and XMP II sets 3200MHz, and the rest of these changes aren't important to me or fine as defaults.

 

Memory Controller DRAM Frequency Ratio: 1:1

DRAM Frequency: DDR4-3200MHz

Intel (VMX) Virtualization Technology: Disabled

Intel SpeedStep: Enabled

VT-d: Disabled

Primary Display: PEG

M.2_1 Link Speed: Gen4

PCIEX16_1 Link Speed: Gen4

SATA Controller(s): Off

Legacy USB Support: Disabled

XHCI Hand-off: Disabled