Is 16Gb of RAM is too low for blender? It’s constantly crashing to desktop, mostly during rendering. Task manager shows 100% ram usage. Should I get some another 16gb module or it doesn’t matter anyway?
Thinkpad P53 with
i7 9850H,
16Gb,
Quadro RTX 3000 6Gb.
Are you rendering on GPU? 16Gb should be plenty for most scenes on CPU (especially if you have swap space available). 6Gb can prove to be woefully inadequate on GPU.
Yes, GPU compute. Are you saying that in my case it’s better to use CPU rendering?
It depends on your scene, but if it will not fit comfortably within your VRAM (and remember, your VRAM also has to house your display), then you will experience problems, in which case CPU is your only option.
You can try optimising your scene (e.g. adaptive subdivision is very memory hungry, so if you can get away with bump/normal instead, do that). Don’t use textures bigger than you need, etc.
I though that for Nvidia, Blender applied a patch that memory buffer wouldn’t crash. Also what exact error is being displayed (usually there should be an error out of GPU memory erorr or something simlar.
If it is just a VRAM issue, then Fully agree with Roken.
A “quick” tutorial that showed up in my youtube list on how to resize materials directly in blender.
ExceptionCode : EXCEPTION_ACCESS_VIOLATION
Exception Address : 0x00007FF67D806D22
Exception Module : blender.exe
Exception Flags : 0x00000000
Exception Parameters : 0x2
Parameters[0] : 0x0000000000000000
Parameters[1] : 0x0000022502B0FFF8
Stack trace:
blender.exe :0x00007FF67D806CF0 MEM_lockfree_freeN
blender.exe :0x00007FF679C01000 ccl::CPUDevice::mem_free
blender.exe :0x00007FF679C16640 ccl::MultiDevice::mem_free
blender.exe :0x00007FF67A085FB0 ccl::shade_background_pixels
blender.exe :0x00007FF67A082270 ccl::LightManager::device_update_background
blender.exe :0x00007FF67A081E30 ccl::LightManager::device_update
blender.exe :0x00007FF679FEEF70 ccl::Scene::device_update
blender.exe :0x00007FF679FF1F50 ccl::Scene::update
blender.exe :0x00007FF67A009A10 ccl::Session::update_scene
blender.exe :0x00007FF67A0080D0 ccl::Session::run_cpu
blender.exe :0x00007FF67A007CF0 ccl::Session::run
blender.exe :0x00007FF67C9DFD10 ccl::thread::run
blender.exe :0x00007FF6798A8350 std::thread::_Invoke<std::tuple<void * __ptr64 (__cdecl*)(void * __ptr64),ccl::thread * __ptr64>,0,
ucrtbase.dll :0x00007FFEA1496BB0 recalloc
KERNEL32.DLL :0x00007FFEA27054D0 BaseThreadInitThunk
ntdll.dll :0x00007FFEA3EA4830 RtlUserThreadStart
MEM_lockfree_freeN
This tells me that there might be no free memory during operation. But anyway, it shouldn’t be doing that this way.
Hi LevAnni,
Here’s a tip that may help with this.
Hope this does the trick.
Cheers and stay safe!
Could you upload a screenshot or two of the scene you’re trying to render? The blend file would be better. 16gb should be more than enough for most scenes. If you have lots of 4k image textures, then that would eat RAM pretty quickly. As does multires and subsurf. But again, there would need to be a lot going on.
@Magnavis Not necessarilly. I can build a scene that looks simple and will eat through 16Gb in no time. It all depends on the content, and how it’s created.
I know this, because I have don it, and then had to go on an optimising mission to get it to renderable. (The last one I got to below 5.5Gb after some work).
There are some textures with 4k resolution, yes, but what I’m trying to say that it should not exit to desktop so unexpectedly.
Nothing special really, basic room scene. What is interesting here is that when it crashes, I’m loading the same scene with same parameters and it renders fine. It doesn’t crash twice at a time, just randomly. This tells me that blender just gives up when it needs more ram, when I’m restarting it uses fresh ram without being filled up during work in viewport. My system is stable, no hardware problems whatsoever.
Hi, I have the same issue since I started messing around with startup files, userprefs and Blender icon: all Blender versions crash on start and I have to use the portable version to open it.
Besides, on this project:
; I have tried to reduce the wight as musch as I could but it still requires about 11Gb of memory.
Any advice is welcome.
Edit: removing the startup file and userprefs solved the crashing issue but I fear that this level of memory will cause many problems for people with average and weaker devices.
