Tag for Blender performance related tasks.
Problem was with Viewport Anti-Aliasing. With it turned off performance is back to rock solid 60, with no perceivable difference in appearance.
Seems an unnecessary option to have enabled by default on Retina machines.
Thu, Sep 24
Mon, Sep 21
Thu, Sep 17
Thanks, great work @Jeroen Bakker (jbakker) !
Wed, Sep 16
Fri, Sep 11
@Sam Van Hulle (sam_vh), This seems to be better in latest master, IMO good enough to close this issue.
Mon, Sep 7
@Demeter Dzadik (Mets), the idea you are suggesting is not impossible to be implemented, but it requires some deep changes in how the system is currently functions. It might also cause worse performance in "real" cases, because it will be more difficult to keep all CPU threads busy bu requiring to wait while tree of animation/drives is evaluated.
Fri, Sep 4
V 2.83.5. Mac OS 10.15.6/3.2 GHz 16 core Intel Xeon W/128GB memory/AMD Radeon Pro 16GB
Tue, Sep 1
Thansk for the report,
Can you try the latest daily build too?
Sat, Aug 29
Compiling FFmpeg with NVIDIA hardware accelerated settings: https://docs.nvidia.com/video-technologies/video-codec-sdk/ffmpeg-with-nvidia-gpu/index.html
Aug 27 2020
There is a new ffmpeg gui which includes options for all the new HDR 10 bit formats and Nvidia based hardware encoding here:
It may be useful to generate test files to detect what codecs currently aren't supported by by the Blender built-in ffmpeg and also compare render times with and without hardware acceleration, in order the check the necessary build flags are being set.
All the new hardware settings in ffmpeg explained: https://developer.nvidia.com/blog/nvidia-ffmpeg-transcoding-guide/#disqus_thread
Aug 20 2020
Currently, the VSE ffmpeg lib can't decode AV1. Ex: https://www.youtube.com/watch?v=PVGeM40dABA
Ffmpeg 4.2 includes a AV1 decoder: https://hub.packtpub.com/ffmpeg-4-2-releases-with-av1-decoding-support-through-libdav1d-decoding-of-hevc-444-content-and-more/
And an AV1 encoder: https://trac.ffmpeg.org/wiki/Encode/AV1
Less than 45 FPS now on 2.83.5.
Aug 19 2020
Aug 18 2020
Aug 14 2020
Aug 11 2020
Aug 10 2020
@Dalai Felinto (dfelinto) Got it! I thought this was exact the same issue, so I didn`t new report. But if U suggest so, I'll try.
Hi @Vyacheslav (hitrpr) - this is already marked as known issue, so although more information can help, the report already has all the information a developer would need to look at it. In your case you can just report your issue as a new bug, and the triaging team can handle the investigation and eventually merge the tasks if it comes to that. Just please refer to this task when opening a new one.
Aug 8 2020
Here is another example. Even if you apply Subsurf, switched Mirror will be calculated on every move. You can test it with this file (if lags are not so noticable — increase subdiv)
Aug 5 2020
Jul 27 2020
D8234: Use DrawManager for Image/UV Editor is more complete, so time to abandon this patch
Jul 22 2020
Jul 21 2020
Jul 17 2020
Added a missing increment of driver_index resulting in incorrect behavior
Jul 10 2020
Jul 9 2020
Huuuuuu.... I did not think this needed to be threadsafe... Does it actually needs to be as such? Linkedlist is not threadsafe either, how comes this is not an issue?
- Removed unused code
- Removed unneeded branch + spelling
- Use runtime flag and runtime evaluation of lookup
Use runtime flag and runtime evaluation of lookup.
Added comments to reset function