Content Aware Fill Performance and CPU Core Usage - Efficiency Cores vs Performance Cores

[sca_75]

Hi,

I have i9 12900KS and this CPU has 8 performance and 8 efficient cores. My GPU is RTX 3090. I`m editing 4K 120 fps HEVC video.

When working with Conten Aware Fill - Subtract operation, if number of keyframes you create are many then the overall process of analysis and rendering takes a very long time. I also tried with a small number of keyframes, analysis took less time but rendering did not take significantly shorter. During this time overall CPU load is around 10%-30% max. GPU utlisation is also around 30%-50% on average. I have 128GB memory installed and After Effects is using between 50 GB- 100 GB of memory when analysing and/or rendering.

I notice that Content Aware Fill is mostly using effcient cores of the CPU - which are much slower than peformance cores - when analysing and rendering. Also my GPU is not 100% utilised during either process. 

Is there a specific reason for using efficient cores for this feature? Is it possible to configure After Effects so that it uses performance cores to improve performance? Any recommendations to improve perfromace is also very much appreciated.

oh - by the way while I was writing this post, After Effects crashed second time before complet'ng rendering 😞 - so this is another bug report that`ll go to Adobe..... 

Many Thanks. 

[ToolfarmJP]

This information might be useful.

https://helpx.adobe.com/premiere-pro/kb/hardware-recommendations.html

[sca_75]

Thanks for the info however I'm already following the guidelines. Please see my hw info in the question. Actually the question is about why the process is using effcient cores instead of performance cores? This could improve the process significantly. 

[OussK]

Regardding the GPY be sure you select the GPU accelerator from the Project settings to mercury GPU as i notice this helps alos in the preview 

[sca_75]

thanks for the info however this option is already selected. The quesiton is actually about using efficient cores instead of performance cores rather than improving the overall performance.