Question Memory issues with new X299 build

[benr_]

I am having relentless MemTest86 errors on my new build. All new components.

Motherboard: Gigabyte X299X Aorus Master (rev 1.0); BIOS: F3c
CPU: Intel Core i9-10900X
CPU Cooler: H100i RGB Platinum
RAM: G.Skill F4-3600C17Q-64GTZKW, Trident Z, DDR4-3600MHz CL17-19-19-39 1.35V, 64GB (4x16GB)
PSU: Corsair AX1600i (because I will be adding more GPUs when the system is stable)
SSD: NVMe M.2 970 EVO Plus
GPU: EVGA Nvidia GTX 1060 (temporary while building)

The machine went together without issues, and booted and installed Windows without trouble.

BIOS is at all default settings, no overclocks. RAM modules are installed in slots per the motherboard manual.

I ran the machine for a day, and installed some typical software I use, also without trouble. I have rebooted the machine many times without issue. Software is all running well from a user point of view.

I wanted to run some proper burn-in tests before calling it a day, and started with MemTest86 running from USB stick.

Immediately on pass 1 of 4 MemTest86 shows errors, and quickly grows to around 100 errors.

I have tested with and without XMP enabled. I also noticed that the Gigabyte BIOS in default XMP "Profile1" selected "auto" for the RAM voltages, and appeared to be setting them to 1.20V, where the RAM spec is 1.35V. I changed that manually to 1.35V, and it didn't help.

I have checked and rechecked every component and cable connection and seating. And as I mentioned, the machine always boots up fast and appears (from a user perspective) to be working fine.

In a crude effort to diagnose the issue, I have now purchased several replacement components (which I will return when I figure out what is going on).

The MemTest86 errors persists with:

• Completely different RAM. I bought and tried a different set of G.Skill RAM (slightly different specs, but also 4x16GB).
• A different copy of the exact same motherboard. I bought a duplicate of the same Gigabyte X299X Aorus Master board, and the same issue persists.

The only time MemTest86 does not report errors is when I use a set of Corsair 2x8GB. Then – with just that 16GB RAM, in just two slots – MemTest86 reports no errors over multiple passes. But long term I'd like my 64GB, and also to know what's happening.

At this point I'm wondering: is it possible I have a bad CPU? It's the only component I haven't changed.

Or am I missing something simple but fundamental in the setup?

I'm not particularly sophisticated when it comes to BIOS settings – I've only built a few machines over the years, and always left most settings at defaults. I don't want to overclock, the machine will be enough for me at stock speeds.

But do I need to change some BIOS settings just to get it stable?

Any advice appreciated. Happy to dig into timings and voltages if anyone thinks that will help.

 

[aigomorla]

bad board or bad cpu, since the problem is persistent when you do quad channel.
Unfortunately the only way to debug from here is if you have a spare board or spare cpu.

 

[benr_]

Thanks, I appreciate the reply!

It can't be bad board, unless I got two identical faulty copies of the board. Is that possible? Likely?

And today I ran the Intel Processor Diagnostic Tool which is a "PASS".

Can the CPU be faulty and still pass the Intel Processor Diagnostic Tool?

Someone else suggested "given that it's a new, cutting-edge CPU, I couldn't rule out some sort of incompatibility." is it possible that the board needs some kind of BIOS update? It doesn't seem so cutting edge components to me…!

 

[aigomorla]

It can't be bad board, unless I got two identical faulty copies of the board. Is that possible? Likely?

And today I ran the Intel Processor Diagnostic Tool which is a "PASS".

Can the CPU be faulty and still pass the Intel Processor Diagnostic Tool?

Someone else suggested "given that it's a new, cutting-edge CPU, I couldn't rule out some sort of incompatibility." is it possible that the board needs some kind of BIOS update? It doesn't seem so cutting edge components to me…!

Did you get both boards new? Or was it am RMA? Can you inspect the pins for any bent pins, these sometimes are known to cause memory errors. If both boards are new and its showing the same issues, its either cpu, or bios related.

Yes, you can have a bad cpu when it passes, because your problem is telling me it is erroring out when the cpu enables quad channel (4 sticks).

That could also be possible that you require bios updates. Did you get the latest bios or does Gigabyte not offer it?


You see how these problems all point out to the board or cpu?
This is why i am still saying its either the board or cpu.
The board could require a new bios tho.
And Gigabyte has one of the worst RMA's. IF the board was a RMA, then i would not underestimate them sending you a bad board as a RMA because the monkeys there still seem to test the board with a 7740X which only uses dual channel, hence why it may have passed QoC and is failing on your end.

As you can see i have very little faith in Gigabyte and RMA's.

 

[benr_]

Both boards were new. I did check both very carefully for bent pins, none apparent.

It's on the latest BIOS, F3c, but I've just been told that likely indicates a beta, and maybe to try rolling back to F2.

And yes, I agree: it has to be something on the board or CPU.

As luck(?) would have it, I am trying right now with a new set of 64GB (2 x 32GB) RAM, and it's passing MemTest86 (from a cold boot).

PassMark support directed me to this thread, which seems to be very much the same issue: https://www.passmark.com/forum/memtest86/46206-help-interpreting-results

So, yes, I agree completely, it seems to be something to do with the quad channel, or (per that thread the 64GB?). I am nervous about re-running MemTest86 after booting Windows first (see that thread). But it's an improvement.

I agree, I think I will likely steer away from Gigabyte in future.

 

[benr_]

After a lot of testing with different components, it turned out there is a BIOS bug in this X299X motherboard.

More information can be found here: https://www.passmark.com/forum/memtest86/46206-gigabyte-x299x-memtest86-memory-errors

My thanks to PassMark support for all their help in diagnosing the issue.

Happily, Gigabyte support are investigating and already sent through one testing BIOS update. That one didn't fix it yet, but I'm happy they're taking it seriously, and hope to have it fully fixed soon.

 

[timatc]

After a lot of testing with different components, it turned out there is a BIOS bug in this X299X motherboard.

More information can be found here: https://www.passmark.com/forum/memtest86/46206-gigabyte-x299x-memtest86-memory-errors

My thanks to PassMark support for all their help in diagnosing the issue.

Happily, Gigabyte support are investigating and already sent through one testing BIOS update. That one didn't fix it yet, but I'm happy they're taking it seriously, and hope to have it fully fixed soon.

I also have this same issue - however with a Gigabyte X299X Designaire 10G board. Did they give you an indication on when this might be fixed?

 

[nat_nrg]

Very similar problems and system here too, @benr_ did you hear back anything else from Gigabyte?

Motherboard: Gigabyte X299X Aorus Master (rev 1.0); BIOS: F3c
CPU: Intel Core i9-10940X
CPU Cooler: NZXT
RAM: Corsair 256Gb
PSU: Corsair AX1200i
SSD: NVMe M.2 970 EVO Plus
GPU: Gigabyte Nvidia GTX 1660 super

I pass MemTest86 but get a hardware fail error on the windows memory diagnostic, a similar pattern to mentioned elsewhere. My crashes mostly occur when CPU utilization is >80%. For example, it will not pass any CPU stress test and will crash within 5 mins despite being within temp limits. Crashes have also occurred in mundane tasks.

Not sure what else to do if Gigabyte cant fix this? RMA and go with a different motherboard?