In it's current form AE will use a portion of a bunch of CPU's but it will not fully utilize all the power that is available. This is simply because AE processes one task at a time and even the largest image frame does not require all of the power available from even a modest system to make calculations on a single frame of video.
Adobe has been hard at work on the underlying structure of AE for a fairly long time but there is still a lot of work to do. Processing video on a frame by frame basis is never going to be able to use all of the power of a modern system because there is simply not enough data that needs to be crunched. On the other hand, even an open source 3D app like Blender has a lot more calculating to do when it's trying to turn some path data into pixels so 3D apps typically will use much more of the CPU power than any app that is processing pixels because the math is a lot more complicated.
AE will run faster on a properly configured powerful machine, but you will not be able to take advantage of all the system resources you have until Adobe (or Final Cut or just about any other pixel processing app) figures out how to process pixels in a more efficient manner.
I use something called BG Renderer that can be found here: BG Renderer Pro - aescripts + aeplugins - aescripts.com
That tool alone has almost halved the time it takes for me to render my animations and also changing the output to multi machine PNG sequence.
it basically queues up multiple render instances on one machine and this is how I utilize 100% of my cpu during the export render process. - Just for some insight.
So to summarize, I believe I can utilize all 90-100% of the cpu cores, as long as AE will recognize them off the bat. So anyone with experience please let me know. Thanks.
It has been more efficient to render image sequences for years, especially when you use a render farm or multiple instances of AE. You cannot render a video with multiple instances because there is no way to keep track of the frames and stitch then together. Your script helps, but doesn't work for all systems and configurations. It is nothing new.
As I said, until the basic way video is processed you cannot force any Adobe or any other video product that I'm currently using to use all available resources to render to a video file. Everything about how video is compiled will have to change before that happens.
1 person found this helpful
Some older versions of AE (CC 2014 was the last one that did it) had a feature called multiprocessing that launched several instances of AE in the background in order to use multiple cores in a system. It worked really well in many cases. There are some projects where it made a HUGE difference on multi-core machines. But, it was fairly buggy and, in many cases, just flat out wouldn't work.
So, the renderer in the new versions doesn't use multiple cores very well, BUT, the software is actually multithreaded now whereas the old versions weren't. The UI and renderer run on separate threads since CC 2015 (but works much better now in the latest few releases). This makes AE much snappier to work with and lays the groundwork for the future improvements that Rick alluded to.
I also should mention that the C4D renderer that comes in the very latest release is actually a fully multithreaded renderer (because it is the C4D renderer under the hood). Also, the new versions have GPU acceleration of several effects.
Keep in mind that a CC subscription includes all versions of the software from CS6 through to the most recent release. That means, you can work in the latest release for the snappier interaction that it has and then render in CC 2014 if your project would benefit from it. But, again, the older versions don't have GPU-accelerated native effects and some expressions or effects are incompatible with multiprocessing so, in lots of cases, the latest version is still faster.
To be perfectly frank, you don't need dual processors or quad processors if you're just doing work with AE in its current form. Higher clock speed is much more important. In the future that could change. And, if you're planning to use Cinema 4D, you would definitely benefit from as many cores as you can get. Actually, multiply the clock speed by the number of cores and THAT'S the number you want to be as high as possible. (Unless, of course, you're planning to use a GPU renderer like Cycles 4d or Octane).
Thank you, This provides a little more info than I had before about the C4D renderer. Though I'm thinking like most things in life, I'm just going to have to learn the hard way and pay to play. But before I do that, I'm going to mess around with some settings maybe I can squeeze a bit more out of my 5960X. Thanks again.
No problem! Glad to help.
Don't forget to mark one of the answers as correct with the green star Correct Answer button to close out the thread.