Giving Up On The Rampage VI For GPU Rendering

[drossxyu]

Is there anyone in this community that has successfully found the right combination of BIOS and settings to get
this board to run without either locking up or getting a BSOD while under heavy GPU load? I've been mostly rendering with
Redshift via Houdini and it's been completely hit or miss. Sometimes I'm running fine for a day. The next day I get 3 hard
lock ups 15 minutes apart.

I'm on BIOS 1704 after flashing it from an equally unstable 15-something

Everything is set to the defaults in the BIOS. I've tried XMP for my ram which it supports but felt like I was getting
further lock-ups. I have not set my RAM parameters manually as of yet.

I've run memtest for 2 passes (all i've had time for) with no issues.

I've tried Prime95 for a couple hours with no issues.

My temperatures are all fine. CPU is water cooled. One 2080 ti is water cooled. The other is 2 PCI-E slots away and has normal temps.

I've yet to try moving my RAM to only the left side which some users have suggested. This is not really a solution to me
as when I decided to upgrade from 64 to 128 this will no longer be a fix.

If someone out there with the same motherboard and similar specs and is doing similar work, I'd love to heard your magical
combination to get this board running without issue.

These are my specs:

Windows 10 Pro 64-Bit
Intel Core i9 7940x @ 3.1 GHZ
64GB G.SKILL Trident Z DDR4-3200MHz
Asus ROG Rampage VI Extreme (LGA 2066)
EVGA RTX 2080ti Hybrid XC
Asus RTX 2070 (blower)
Samsung SSD 970 Pro 1TB (M.2)

[LiveOrDie]

Guessing your CPU is running stock? maybe try overclocking it just saying sometime stock can be worse than a good solid OC.

[drossxyu]

LiveOrDie wrote:
Guessing your CPU is running stock? maybe try overclocking it just saying sometime stock can be worse than a good solid OC.

If it's not overlocked when set to the default settings, then no it's not. Is there some indication that makes you think the CPU clock speed is causing the instability? Genuinely
curious.

Do you have a recommendation as to what means I should OC the CPU in the BIOS. I am not that familiar with the intricacies as I bought this board because I thought
a board built for OC would be the most stable for someone that had no intention of overclocking.

[drossxyu]

drossxyu wrote:
If it's not overlocked when set to the default settings, then no it's not. Is there some indication that makes you think the CPU clock speed is causing the instability? Genuinely
curious.

Do you have a recommendation as to what means I should OC the CPU in the BIOS. I am not that familiar with the intricacies as I bought this board because I thought
a board built for OC would be the most stable for someone that had no intention of overclocking.

So no one else is able to share their experience getting this board running stable? *Has everyone else dumped it for something else?

[LiveOrDie]

drossxyu wrote:
If it's not overlocked when set to the default settings, then no it's not. Is there some indication that makes you think the CPU clock speed is causing the instability? Genuinely
curious.

Do you have a recommendation as to what means I should OC the CPU in the BIOS. I am not that familiar with the intricacies as I bought this board because I thought
a board built for OC would be the most stable for someone that had no intention of overclocking.

Just having a guess on experience here also have you tested to see if your CPU hits the turbo boost clocks it should? sometimes a combo of components need to be fine tuned you might have 1 lazy core or a weak IMC, does this happen when running no XMP? and just because a board is made for overclocking it doesn't mean its more stable the amount of features they cram onto these boards sometime backfires so most of the time a basic board will be more stable overall.

[Hopper64]

What's the exact part number for the memory? Have you tested it? Is the kit on the QVL list? This is first few things I can think of at the moment. Ramtest is recommended by a few on the forum as it's much quicker:

https://www.karhusoftware.com/ramtest/

[drossxyu]

G.SKILL TridentZ Series 64GB (4 x 16GB) 288-Pin DDR4 SDRAM DDR4 3200 (PC4 25600) Intel Z370 Platform / Intel X99 Platform Desktop Memory Model F4-3200C14Q-64GTZSW

I'll give that test a try. Thanks!

Another thing I just tried was using the OC wizard in the BIOS just for the hell of it. The first OC'd the clockspeed 35% and the second 20%. Instant freezes upon entering windows. WHEA_UNCORRECTABLE_ERROR. I have an NZXT Kraken x62 on my CPU which I forgot to mention. I had to end up doing a full CMOS reset to get it running without instant lock ups. Just setting it back to the optimized defaults was not enough to bring back the "stability".

Here's a post I came across that eludes to the same issues as mine while rendering:

https://render.otoy.com/forum/viewtopic.php?f=30&t=69144&p=349896&hilit=rampage#p349896

There seems to be a fix here but is beyond my level of comfort.

[drossxyu]

And to answer your QVL question. Yes, my ram is on the list if your search for "F4-3200C14Q-64GTZSW".

Running RAM Test right now for the past 15 min. Error Count 0. I'll run it for a longer duration later tonight and get back to you.

[Hopper64]

drossxyu wrote:
And to answer your QVL question. Yes, my ram is on the list if your search for "F4-3200C14Q-64GTZSW".

Running RAM Test right now for the past 15 min. Error Count 0. I'll run it for a longer duration later tonight and get back to you.

That's reassuring. You might want to post pics of your bios settings for everyone to look at.

[drossxyu]

Ran it for 49 min reaching 927% coverage. 0 Errors.

Despite this I opened up CPU-Z and noticed my DRAM Freq and timings are wrong.

They should be :

~1600mhz (half of 3200)
14-14-14-34-2N

Instead it shows:

~1065mhz
15-15-15-36-1T

So I figured why not enable XMP to fix this. Go into the BIOS (which btw is absolute defaults at this point) and flip on XMP. I deny whatever intel optimization it tries to turn on when you enable XMP. Boot up. Confirm Memory tab in CPU-Z looks accurate now. Fire up Ram Test for about 10 min. Locks up. No BSOD or windows error. Just frozen.

So to recap what's happened this evening:

Do some basic OC using the BIOS's built-in OC wizard tool : WHEA_UNCORRECTABLE_ERROR
Set my ram specifications correctly via enabling XMP : Freeze up within 10 min of Ram Test.

I'd share an image of my BIOS but I have reverted to defaults yet again. Everything seems to be on AUTO.