21
votes
Why/how does tile size effect performance?
The difference and why there are different optimal bucket sizes comes from the size and design of on chip cache/memory of cpu or gpu.
The gpu has a massive amount of cores. But they are dumb and ...
17
votes
How to make VSE render faster?
Update:
Multithreaded VSE rendering is now natively implemented in blender (since around 2.8 or 2.9), which makes this project kind of obsolete, but for more fine grained control of resources (...
17
votes
Accepted
What's the difference between the various Open Subdiv compute types?
The difference relies in in which device the actual calculations are done. CPU and OpenMP are run on the CPU, while GLSL Transform Feedback and GLSL Compute are run on the GPU.
CPU - single ...
15
votes
What computer should I buy for Blender? How do I choose hardware?
According to rules, hardware questions are off topic here, you can ask about them over at Hardware Recommendations or other third party communities like Blender Artists Forum. To not leave you empty ...
14
votes
Accepted
Why is my cycles render still grainy on 1000 cycles?
I see a few things you can improve:
You can use much less polygons. The higher the polygon count, the longer the render will take. Try searching for tutorials that use subdivision ("subdiv") ...
14
votes
What are the pros and cons of using Blender on Windows or Linux?
It is commonly said that performance is generally faster under Linux based operating systems, both during regular program operation and rendering.
Linux is also commonly said to consume less resources ...
14
votes
How to find the number of loose parts with Blender's Python API?
Recursive Bmesh version
Similarly to the way other bmesh operators, BMesh.Ops are used,
get_islands(bm, verts=[])
...
13
votes
Accepted
How to find the number of loose parts with Blender's Python API?
I propose another approach based on dictionaries and sets, for performance purpose. It allows here to calculate a 1080 times arrayed sphere with about 500k vertices in 3 seconds.
The principle is to:
...
12
votes
Accepted
What is Clamp Direct/Indirect?
Clamp puts a limit to the intensity of light bounced directly and indirectly by surfaces.
It can be used to prevent the annoyoing "fireflies" (random intense pixels in your render) but it is advised ...
12
votes
Is it possible to render a movie on a single computer in blender?
It is possible to render a movie on a single computer. It depends on what movie it is. If you choose a style that renders extremely fast, it is possible. Nothing like the ones Disney make though. Not ...
10
votes
What are the pros and cons of using Blender on Windows or Linux?
If you are using CUDA rendering, Linux is faster by 20-40% due to the overhead of WDDM in Windows 8/10. This is regardless of if there is a display plugged into the GPU or not.
To work around the ...
10
votes
Accepted
Should I wait when Blender isn't responding?
You can wait if you want, but it's probably going to take a long time before it will start responding. If you close it there is a way to recover you data; Blender autosaves your projects every 5 ...
10
votes
Accepted
Blender Global undo is very slow on complex scenes
The undo buffer works by actually saving the entire blend file in memory, and loading it to undo. If you have such big, complex files, that'll take a while.
A possible way to speed things up would be ...
10
votes
Accepted
How to display calculation time above nodes?
To do this, simply activate the "Timings" option in the top right menu.
Note, however, that this is more of a guideline than a reliable indication of the computing time.
Changing the ...
9
votes
Accepted
Are there any ways to speed up blender compositor?
You can change the lines to be straight by going to "File > User Preferences > Themes > Node Editor > Noodle Curving".
You can disable "Use Nodes" to keep it from updating until you are done moving.
9
votes
How to find the number of loose parts with Blender's Python API?
That code makes a different thing - it finds the vertices that belong to the loose parts. But, the code is much clearer. If you need the count, just take len().
...
8
votes
Accepted
OpenCL vs CUDA Rendering
Go with the Nvidia. Performance aside, CUDA works better with Cycles. Until recently, OpenCL was not even supported in Cycles so it is not as bug-free and stable as CUDA. Another thing to consider ...
8
votes
How is real-time rendering different in blender from game engines?
Well, Blender isn't a game engine, and it's not designed to perform final animation renders in real-time the way a game engine does. The design goal of the EEVEE renderer is to provide real-time or ...
8
votes
Need Technique for Projecting Countries onto the Earth
You could:
Select all you countries and join them with CtrlJ
Create 2 grids (delete faces only, keep the edges), one for the front, the other for the side view:
Select one grid, shift select the ...
8
votes
Geometry nodes vs particle system
The big deal with geometry nodes is that we're effectively getting the functionality of geometry shaders now. That's kind of a blunt way of putting it, but if you look that up in the context of GLSL ...
7
votes
Accepted
Render FPS is low
Slow FPS is a very relative term. Considering the textures you have and your rig, I think 5 FPS quite a good performance rate.
Note that Blender is not heavily optimized for real-time graphics like ...
7
votes
Accepted
How to disable entire tracks (or multiple clips) in VSE
To disable within the VSE, use Mute Strip.
Mute Strips: H
Un-Mute Strips: Alt + H
7
votes
Geonodes: which is faster, Set Position or Transform node?
Blender 3.6.0
Testing on Suzanne, ✲ Ctrl5 5 subdivision levels, and this setup:
All measurements are done on the Group Output of the above, while modifying the contents of the i custom group. ...
6
votes
Accepted
Faster blur effect in Blender VSE
Since I figured out the answer, I'll answer my own question.
Blender's Video Sequence Editor is an awesome tool; however, it lacks some functionality like some effects the Compositor has.
For ...
6
votes
Optimize cycles background rendering
As of Blender 2.83 (currently in alpha at the time of this post), this can be solved by using the "Adaptive Sampling" option in Cycles. Adaptive sampling will automatically stop sampling in areas of ...
6
votes
Why/how does tile size effect performance?
One thing not mentioned that is important for efficiency on multi-core hyperthreaded machines is the time cost of "tile stragglers." My machine has 12 cores and will spawn 24 rendering threads which ...
6
votes
Accepted
Render using all cores except one (or N)
You will have to have a script file in each of your blends containing this:
...
6
votes
Accepted
Multi-core/threaded particle simulations
In Blender particles and simulations are still badly optimized and mostly single-threaded with no GPU compute support. With time features are added and even now you can look into the FLIP FLUIDS addon....
Only top scored, non community-wiki answers of a minimum length are eligible
Related Tags
performance × 457rendering × 123
cycles-render-engine × 84
gpu × 50
3d-view × 43
animation × 30
hardware × 26
python × 23
modeling × 18
interface × 16
eevee-render-engine × 16
windows × 16
optimization × 16
video-sequence-editor × 15
scripting × 12
memory-management × 12
modifiers × 11
objects × 11
particles × 11
edit-mode × 11
materials × 10
game-engine × 9
workflow × 9
sculpting × 9
mesh × 8