Hi all, I’ve been getting these ever since i’ve upgraded to 5.1.1 (might be a coincidence with the update) and I can’t seem to get rid of them.
They first appeared when i activated Nanite for a static mesh, and even though i’ve deactivated it ever since, they still show up in the editor and PIE.
I have upgraded my graphics drivers, to no avail.
I’m wondering if anyone knows anything about this
Thank you
I think you’re running out of VRAM. If you restart the editor, they go away, right?
1 Like
yes, but they come back pretty fast next time i open the editor. I have a 3070 with 8gb of VRAM, and the scene isn’t that complex, do you think it could be faulty scene optimization or a problem with the video card?
- also, i’m not getting texture streaming pool errors or anything, or, rather, when i do get them, they’re don’t seem to be correlated to this issue
Texture streaming is not the same problem, you would actually get VRAM messages.
I regularly run into the problem in one of two circumstances
-
A lot of textures in the scene
-
I have been auditioning a lot of textures during the editing session
From the sound of it, for you it’s 1.
1 Like
this can’t be it, i have less textures in my content folder than i do in the engine one. It’s just a couple of quixel surfaces repeated everywhere.
I think it has something to do with Nanite, the whole thing appeared when i activated nanite on one static mesh and then it kept on happening even after i turned nanite off.
Do you know if nanite also does stuff to the textures?
OK, you’re right 
I don’t know enough about Nanite yet, sorry.
I posted a solution here that should fix the issue
Hey there @Evilnescu! Texture resolution plays a large part in this as well. You’re DX12 and SM6 because you’re using Nanite, so you’re going to get far more overhead than most projects.
Are these textures 4k - 8k?
Nanite technically should reduce GPU overhead, not create it usually. This can sometimes stem from Nvidia’s settings configuration or older. Techtut’s solution might work out immediately if this one’s the case.
Is this accompanied by crashes? Particularly the DX3D driver failing or Rendercore crashing out?
Sorry for the late reply, i’ve been away on holiday.
not more than usual, i generally restart the engine before a crash would happen, because it’s really unnerving to look at.
thank you, but i don’t know what to do with that code, where to put it? in the console?
All good, hope you enjoyed it! So as Clockwork mentioned this can be associated with your GPU or it’s drivers failing one way or another. If your textures are too large you’ll run into this issue often, other times it could be your mipmap levels on the large textures being too fine and gobbling up VRAM.
Are you using anything like iCue, MSIafterburner, Shadowplay, or any other 3rd party apps that touch the GPU in any way?
i still have the issue and i still don’t know what to do.
I don’t use any 3rd party app for the GPU just the standard driver from nvidia, updated properly ofc
Hrm, have we checked to see how much GPU utilization is going on? Is this still occurring on less populated maps? Check out this thread where some GPU shenanigans were happening.
what do you mean by “change the texture group”?
Like for every object? The glitch happens to random objects in the scene, regardless of their materials etc.
To me it seems like it’s a systemic issue, having to do with the render pipeline.
Also definitely can confirm it’s something that resets with an engine restart, but like i said, it shows up faster after the issue happens - which indicates a driver/videocard issue.
I definitely agree it’s GPU based, but if your card isn’t hitting it’s head on the cap, you aren’t getting VRAM or Texture cap issues, it could just be that the drivers are failing at higher utilization but that’s a hard call to make from just artifacting. Let’s see if it’s just heavy utilization, then we can lean into more drastic fixes.
Drop the scalability to low/medium and keep it there for an entire session and see if it takes much longer for this to appear. After that, try working until this appears, pop open another default scene and just place a couple of your meshes in and around to see if the memory ever gets released.
I definetly think its engine based. I tested it with an RTX A6000 (drivers are updated) and bumped up the VRAM Pool of the engine to the max - guess what? - textures still freaking out
Interesting, great data! Most times when I see this specific texture artifacting in an otherwise functional (while heavier) scene, it’s either the scene is too heavy for the renderer and instructions are getting botched coming out to the GPU or the instructions are being misinterpreted by the drivers once they get there. Hard to find a fix in this case, especially if it’s engine limitations. Since this is mostly quixel/mixer stuff it should be all Master materials and instances so it shouldn’t be optimization.
Think this is a UE5.1 thing.
Same Project, Content, etc → UE4.27 and UE5 are working fine with the textures
1 Like