New user. RTX 3080 card and crashes
davorama_9f606bd07d
Posts: 50
I just downloaded and installed the latest version of DAZ STUDIO. I seem to be having an issue with the iray preview mode. When I use it for a while, my whole computer locks up, especially if I'm adjusting materials. I do use the nvidia studio drivers and the latest drivers are installed. Anybody else have problems with this? I have to hard shut down my whole pc, it doesn't just affect DS.
Cheers,
Davo
Comments
Download gpu-z and check if the temps are ok.
Could be a defect graphics card, or maybe your power supply is too weak?
Depending on the model the 3080 can draw up to 450 watts.
What other hardware do you have in your system?
Does it also happen in other applications like a gpu intensive benchmark?
I have a 3070ti with a 700 watt power supply and never have issues.
What are your system specs, how much RAM do you have?
Go to Render Settings, click on the Advanced tab. Under Devices for both Photoreal Mode and Interactive Mode, is “CPU” checked? If so, uncheck it. Under General, is Allow CPU Fallback checked? If so, uncheck it. When CPU is checked, it will put the bulk of the rendering load on your CPU, and that will shoot your CPU temps up rapidly and can cause a Blue Screen. Your 3080 can do a better, faster, and cooler job.
There is no reason to uncheck "Allow CPU fallback", in fact it may result in unforeseen problems (rendering starting over and over again)
If you leave it checked and your render exceeds the memory your video card has available, DAZ will try to render using the CPU instead. Hence the CPU fallback. Which will quickly shoot your CPU temps up to 100° or so, fill your room with the heady arouma of frying electronics, and potentially freeze your system up or fry your CPU. With the box unchecked, if you exceed the memory the render will just stop, and you then can decide what to remove from the scene (or buy an RTX 3090). I keep mine unchecked, and have never had a problem.
A well made machine should handle CPU renders without any problems at all (other than being slower than GPU ones). I CPU rendered for a couple of years before I purchased my first Nvidia RTX card. I never had a problem with the machine, and my dogs got lots of walks while I waited for the renders to finish.
https://www.daz3d.com/forums/discussion/562121/itirations-keep-resetting-to-1#latest
Good info on the AI denoiser , thanks.
The crux of the matter though is CPU temps. If you trust that your chip will throttle performance to keep from exceeding the Tjunction, and aren’t worried that extended periods of running temps right at that limit won’t hurt anything, then by all means CPU render. I like to err on the side of caution, and have learned to live within the means of my GPU’s memory. Either way, I’d bet a nickel that the OP’s mystery problem comes from CPU overheating, in which case he needs to uncheck all three boxes to put things right.
I have CPU rednered on a variety of machines and never seen 100C.
The issue is ocurring in IRay preview only, not during rendering? Is anything else trying to use your GPU?
We still don't have the amount of installed RAM on OP's computer, is it 4GB, 8GB, 16GB, 32GB or more?
Apparently this isn't an uncommon issue with the RTX 3080.
No, not all of them are broke.But a large number of them were shipped with questionable capacitors.
https://techraptor.net/gaming/news/nvidia-rtx-3080-cards-are-having-crashing-issues
Google/duckduckgo/bing or whatever you use for search should find a few thousand related hits.
Here's a thread started on the Nvidia forums back in 2020 about it.
https://www.nvidia.com/en-us/geforce/forums/geforce-graphics-cards/5/405129/3080-random-crashes-and-black-screen/
Going to need to find out what 3080 you have OP and see if it has the defective capacitors or if it's something as simple as the GPU isn't seated in the socket correctly.
If you are running any kind of overclocking software I would revert the settings to default and see if your card is more stable.
----
If you are seeing 100C CPU temperatures , you have cooling problem that needs to addressed.
Well, I'm convinced it's the graphics card. I had another crash and this time the pc won't restart, put in old gtx card and it started up, took it out, started up again with the rtx3080. While I had the gtx card it, I had no issues with iray preview mode. To answer a previous question, it was during preview mode that these crashes were happening. On the crash that I could not get the pc restarted, it happed during a simple iray render, which is a first.
Here are my system specs: (see attached).
On a side note, and kind of a surprise to me, the graphics card info says it's a Dell oem, which suddenly doesn't make me feel confident..
https://www.techpowerup.com/gpu-specs/dell-rtx-3080-lhr-oem.b9508
Not sure
oops, forgot to attach the file. See attached.
The hottest I see in the crash image is the VRAM, and that is not worryingly hot at all. If it gets into the 90's then it is reaching the too hot realm. 65 for the GPU itself, that is really cool under load. I have a huge case with a ton of fans on it, so airflow is great and my 2080 super and 3070ti it's not unusual to see it hit 80's under load, even with the triple fans they have.
My 3080 card made the screen go black while I was running rhino and didn't want to restart the pc. I put the GTX 750ti card back in, it seems to be running stable. I'll have to contact dell and have a discussion about what to do. Hopefully it's just a bad card/fluke. I have been reading up on the LHR (limited hash rate) cards and it seems it's an invidia thing in their new cards, but it should not affect rendering, just crypto mining. I'm not sure who makes the Dell OEM cards but I believe it's MSI.
Your DS preview mode is using thr RTX 3080. Is Rhino also trying to use it?
You are running an i9-12900K. Intel CPUs draw a lot pf power and run hot. An RTX 3080 also draws a lot of power and runs hot, much more so than the GTX 750ti. Is your CPU liquid cooled. How many fans in your case? 77C (170f) is very hot. Could be power draw, as not all 750w power supplies are created equal, what king do you have?
2 case fans, 2 fans built into gpu, power supply has fan. Not liquid cooled. Not sure what brand power supply. I have a ticket in with Dell, we'll see what they say.
What kind of Dell is it - Alienware, XPS, ...?
As a test, I recommend trying to undervolt your 3080 in case its overloading your power supply. This is a common issue with certain 3080's that only shut the system down when playing certain games or 3D applications. Here is a YouTube guide you can try:
I have a 750W power supply running my 3080 GPU. Most of the time its fine, but certain games will force my system to instantly shutdown. While my power supply is 750W, it is also 7 years old and has probably degraded over the years. Using this undervolt method prevents my system from freezing with little to no performance loss. If you use MSI Afterburner, you can save multiple profiles. When using Daz Studio, switch to the undervolted profile, then switch back to a default profile when done.
Also make sure your PC has circulation at the bottom underneath the case. If the power supply is located at the bottom of the case, it sucks air from the bottom to cool it. If your PC case is on carpet for example, you might be blocking crucial airflow from the bottom causing the power supply to overheat. If your PC is on carpet, you can try placing a flat board like a shelf undereath it
It's an XPS 8950. It's power supply is at the bottom of the case, however the case has feet so there's about an inch of clearance beneath. I have room for one more 5" fan, midpoint/near top of front of case. The system is only 3 months old.
A Dell system should be pretty solid. Best to pursue a warranty claim.
What happens if you just run DS with Iray preview, and no other 3D or CAD software? Does it still crash? What was Rhino doing when your system crashed? Was it rendering, simulating, etc.?
I hope your problem is solved. There is nothing like a good warranty!!
Thank you all for your comments and suggestions! My replacement GPU got here today and seems to be running stable (so far, fingers crossed).
As a side note, a little history on this new rig. I got this Dell XPS 8950 system in February and right off the bat, I could not get it to properly go into sleep mode. Screen would go off as specified, then the system would sleep, but it would wake itself up every few minutes and the lighting on the GPU would never go off. The power switch would flash a code and go orange, I had to hard reboot the system to get it out of that error. Dell sent a new motherboard (installed by tech, part of warranty). Didn't solve the issue. Send the whole pc to dell and they reinstalled OS, nothing else done. When I got it back, it still wouldn't go into sleep mode properly, but at least it didn't flash any errors. I just set it to never sleep, other than the monitor. I always wondered if it was a GPU related issue, even the tech guy said "hmmmm maybe". My new GPU allows the whole pc to go into proper sleep mode, light on the gpu goes as well and the HD, all of it powers down. This could all have been a GPU issue the whole time.
I did install an addition fan, so additional cooling.
This is the first Dell rig of many that I've had an issue with, and I know there are a lot of Dell horror stories, but Dell has been helpful and prompt the whole time, sending parts overnight through FedEx, free return shipping for parts I have to send back etc. My only complaint working with them is through chat, I know that they are foriegners and the reading off of que cards is so time consuming, I want to get to the meat of the matter.
Thanks again everybody!
Davo