I suspect most effects aren't well parallelized to make use of multiple cores, so adding more wont expedite the render times much.
Thanks but what about the extra ram? Going from 48GB to 128GB? Also he said the scrubbing wasn't much better either.
Well, straight up: You don't seem to understand how AE works. You can enforce nothing in this crooked little program. It takes what resources its code allows it too and since a lot of that is barely multithreaded, beyond a certain point you will not see any gains. AE simply doesn't use extra cores, because many algorithms that deal with video are linear, encluding encoding to clip based files. The same goes for many effects that need to be temporally consistent based on linear time. The rest is just AE being an almost 25 year old program and still carrying a lot of legacy around in terms of how things are done. You can't solve any of these things with hardware.
You are correct, I don't know how AE works. I am in IT so that is my ignorance. I appreciate the explanation. So is there nothing that would make it run faster other than maybe a new tower with a faster processor or is that not a big gain either?
Appreciate all the info!
The cache drives...do you mean the drive AE is using for its cache or do you mean the place where footage/assets are stored and the place to which renders are sent? AE's cache prefers a fast drive that is separate from footage/renders, so AE's cache should be set to the fastest drive on the system. I have two SSDs in my production machines, one for the OS/Apps and one dedicated for AE's cache. In my freelance studio, I have normal hard drives for footage, etc. (And at my full-time job, it's a fiber-connected SAN for the footage/renders, but still - similar principle.)
The multiple processing cores will help if the artist is using the Cinema 4D renderer to create extruded geometry because it is fully multithreaded. But that only applies to rendering the extruded geometry and not to anything else.
The extra RAM probably won't help either as far as speed goes since you already had so much to start with.
How fast is the clock speed in the system now vs. what you might upgrade to?
Now, what you can try is have the user work in CC 2017 (faster interactivity) and then save a copy to AE version 13, open it in CC 2014, make sure multiprocessing is enabled in the preferences (with 1/4 of the RAM reserved for other programs), and try rendering it out of AE CC 2014's render queue. The old architecture in CC 2014 had this thing called multiprocessing that spun up other instances of AE in the background to utilize multiple processors. It had its bugs (thus the move to the new architecture that isn't quite complete yet), but it may work for your case.
Of note: each new release of AE comes with more and more GPU acceleration of native effects, so hopefully the next release will have some more in there. I've noticed a massive difference rendering the Lumetri effect (for example) on GPU vs. CPU.
Thanks Szalam for the info! I have a separate cache drive but it's a 4TB 7200 RPM drive. The OS/Apps drive is a 1TB SSD drive. Not sure how much Cinema 4D rendering he does. The clock speed is going from 3.33 GHz to 3.46GHz so not much as far as single core is concerned. I've read about the 2014 being able to take advantages of multiprocessors but there are a few people involved and saving down will not be an option since they will not do it. It was not a problem when we were on 2015 but once 2017 came along, that wasn't an option.
Hopefully AE will start to take advantage of multi cores one day. I did changed them from CPU to GPU and they did say it was a big help.