Massive Crashes on two systems (2080ti) Tweet

[kylestauffer]

Hey guys- been fighting this for months and have to find a fix. Here's a quick backstory:

I had a monster PC built with 8x 1080ti's from Velocity Micro. We can't get it to render anything other than a few frames before crashing out to the blue screen and locking up. We even had the tech that built it fly on-site and check every piece of hardware, put in new processors, etc. No luck. That system is running C4D R18 and Octane 4 RC2.

THEN I had a new box built that just arrived this week thinking it was the computer. It's a PC with 2x 2080ti's in it. AND ITS CRASHING THE SAME WAY. I've tried C4D R19 and R20, with Octane 4 all the way to RC5, and even back to the last stable 3.08.3 release. I'm really at a loss. I've tried to enable logs to see what's going on but it doesn't seem to be saving any data on the crashes. The blue screen error is an uncorrectable hardware error, and it never even reboots. It just bricks.

These are both Windows 10 boxes, and admittedly we're entirely mac based otherwise, so I don't even know how to "tinker". This computer is literally fresh from the factory with new install of only C4D and Octane.

Happy to pull files or anything I can to help one of you Octane geniuses on here shed some light on what might be doing this.

[frankmci]

We are also mostly a Mac house with a few PC render machines. I'm guessing you're dealing with a driver issue. In my experience, Windows OS installers and updaters don't do diddly for your graphics drivers. They will be perfectly happy completely ignoring installed but out of date/incompatible drivers without ever mentioning the fact. You will have to take care of updating/troubleshooting them manually.

The first thing I'd try is pulling all but one of the graphics cards from your rig, and testing for stability with the latest drivers. If that works reliably, try it with two cards, etc. If things are unstable, start stepping back through previous driver releases.

There's no guarantee that the most recent drivers will be the fastest or most stable.

Yes, it's a colossal pain in the bum.

[Joss]

https://forums.guru3d.com/threads/windo ... ts.378044/
https://www.redshift3d.com/forums/viewthread/20408/

That might be a PLX chipset problem.

Or some other problems described here:
https://www.redshift3d.com/forums/viewthread/14488

[BCres]

Kyle,

Sorry to hear about your problems. I build my own PCs so I'll take a shot even though these things can be impossible to diagnose remotely. Your system builder has problem done most of this.

Your 1080ti box is configured much like mine so let's start there. Could it be a power supply issue? Anything 850 or higher should suffice for your set up.

I agree with frankmci...pull one of the cards to test. (Make sure you disconnect the power from the wall and let the motherboard discharge fully before yanking the card or you can damage your motherboard.) Then swap and put the other one in by itself if you still have problems.

Much like Joss suggested, get the motherboard info and check the net for issues related to that particular one. I stopped using Asus years ago because of quality issues.

Next, if you know how to get into the BIOS, do so and turn off all overclocking if it's turned on. This can cause a lot of problems just by itself.

Now software...

There should be a "This PC" icon on your desktop. Right click and open "Properties", you'll get a Windows 10 information screen. On the right open "Advanced System settings" then click the "Hardware" tab and then the "Device Installation settings" make sure "no" is selected. Otherwise, Microsoft will put in the hardware drivers they have on their site which are usually out of date.

Then go out to Nvidia, download and install the newest video card drivers. No need to install Geforce Experience or the audio drivers unless you need them.Choose that under Custom install.

Once installed you will have an Nvidia Control panel right click entry on the desktop. Open it and under help there is a "system information" entry select it then components. There will be listed all the versions including grpahic driver and Cuda. You can see if you're up to date.

Also, right click on the Windows button and open your device manager and see if you have any yellow exclamation point on anything else that would indicate another hardware/driver issue.

About all I can suggest right now. I'll watch this and see if you have any other questions.

[glimpse]

Good day.

please list Your components out here piece by piece, also add any other relevant information starting from OS, installed updates, driver version, etc.

I'm pretty sure there is a solution to this problem, but without knowing more info on this it would be hard to help.

Sinc.
tom

[hausgross]

Hi!

Please let us know, what the message of the blue screen is.

When you are having problems after updating to the newest Nvidia drivers make sure to test some earlier versions. They might be less ideal for games, but more stable for Octane.

Also make sure to disable Nvidia SLI in Nvidia Control Panel or even remove the physical hardware bridge connecting the cards. Octane doesn't need SLI and it leads to crashes.

[kylestauffer]

Hey guys- Thanks for all the help so far! Here is the system spec on the second (dual 2080ti) system that i'm trying to get working first:

System: Velocity Micro ProMagix HD80
Processor: Intel Core i9-7920X CPU @ 2.90GHz
Memory: 64GB Ram
Video Cards: Nvidia GeForce RTX 11GB EVGA 2080ti (2x)
OS: Windows 10 Home
OS Version: 1809
OS Build: 17763.55
Power Supply: 1300 watt EVGA SuperNova
Motherboard: ASUS X299 WS Sage Intel chipset

Current Software:
Octane: Octane 4 RC5
Cinema 4D: R19 & R20
NVidia Driver: 25.21.14.1634
NVidia Version 416.34

[kylestauffer]

Hi!

Please let us know, what the message of the blue screen is.

When you are having problems after updating to the newest Nvidia drivers make sure to test some earlier versions. They might be less ideal for games, but more stable for Octane.

Also make sure to disable Nvidia SLI in Nvidia Control Panel or even remove the physical hardware bridge connecting the cards. Octane doesn't need SLI and it leads to crashes.

The BSOD is a generic "WHEA_Uncorrectable_Error" message that just sits there and never restarts. Same error on the smaller PC as is on the larger box. That's a hardware error but no details at all, i'm assuming video card / motherboard related of some sort based on what's been shared here so far.

[glimpse]

try this please: https://windowsreport.com/fix-whea_unco ... ndows-8-1/
(even if it is for 8.1, could be the same on W10)

Overall, think this is totally solvable, I hear about it first time here to be honest ;)
so I doubt it would be in any case related to OctaneRender itself.

Also You have pretty good hardware & I know some guys who have same mobo, couple of 2080tis. It seems to be rather stable.

[DinoMuhic]

Hey I'm having nearly identical setup (just the 7940x for difference).

Got 3 times a BSoD with RC6 and RC7 in the last week. Also whea_uncorrectable_error 0x124

However when I used RC5 for rendering a week prior everything worked.

I did a lot of troubleshooting and yesterday I was not able to reproduce the BSoD anymore. However I did TWO things, not knowing which of them did the trick (if it is indeed fixed, I'll monitor it closely, I hope it is fixed):

1. Turn off Turbo Boost in the Intel application that you probably installed. Its called "Intel TBMT 3.0 Control Panel"
2. Use the MSI-utility fix that is being used for PLX_watchdog violations on PLX boards.

open the program, look for your GPUs, turn on the checkbox for the first one (not both). Hit apply in the upper left corner. Close the app and restart your machine. Then do the same for the 2nd GPU. Again restart the machine.

Usually the BSoD happened while the IPR was on in Cinema and rendering. Like I said, after I did the two things up there I was able to work for hours without nothing happening, I hope it is fixed.

Now another things comes to my mind:
Do you have the ASUS AI Suite installed? It all started for me when I played with the CPU Optimization features inside the app. I had to completely get rid of the software, cleaning registry and everything, then reflash the BIOS afterwards (version 601 I think). Its a piece of crap software and really hard to get rid off.

If I still keep having the problems, I will do a windows clean install next. Because I believe its something with the settings and not really a hardware issue.