You CPU is fine!
RAM... For AE, having your current 16GB of RAM is definitely a limiting factor. 32 or 64GB would definitely help, but I'm not sure about your idea to have 16GB + 32GB (48GB). Also on the subject of RAM, 1600MHz works pretty well; possibly you could afford 64GB of that speed?
GPU... The GTX 780 would be better than the 770GB 4GB for ray tracing speed. While AE is RAM thirsty, it does not seem to need VRAM to tap the ray-tracing calculations.
Thanks for the quick response Jim! Would you happen to have a link to any studies you found that show that VRAM quantity isn't very important?
I guess I'm really just curious about how AE uses the VRAM and CUDA cores. Like what exactly is stored in VRAM and what isn't. And why does adding more CUDA cores eventually not matter anymore? Is it because something else in the system becomes a bottleneck which prevents the GPU from reaching it's full potential or is it because of how AE uses the GPU that it just doesn't matter after a certain point?
I'd also like to know the same about how the Element 3D plugin uses the GPU, but that's a question for a different forum... Unless I get really lucky and someone who happens to know that answer reads this post
I don't have the link handy, but Google Teddy Gage After Effects benchmark and you wil likely find reports of Ray-tracing using After Effects for various CPUs and GPUs.
If is good to be curious and ask questions, however there is nothing like using YOUR software, YOUR workflow, and YOUR media and projects to see where your bottlenecks are.
Simply do various steps and use available CPU, RAM, GPU, VRAM, and thermal monitoring applications to flush out where YOUR system is weak. Then, address those specific issues.
Don't forget setup too; there are lots of places where RAM, cache, swap, and other items can be configured to tune and tweak your system for your workflow.
Element 3D requires a card with 1GB of vram or more. That means ideally you want a card with 2GB of ram or more. The Acceleration with AE is far more limited as to when it's used so I would not worry so much there. Ideally a 770GTX 4GB card would be the way to go without spending the price on the 780Ti.
Thanks Jim, that's great advice for determining bottlenecks. Unfortunately I've already determined that the video card is my next bottle neck so now it's just a matter of figuring out whether to get more VRAM or more CUDA cores.
I also found the Teddy Gage benchmarks which were a little helpful. It was a bit hard to compare the video cards though because the processors also changed drastically. Thanks so much for all the help!
Thanks for the link! That definitely helped shine some light on a few things and helped me figure out that I am one of the people who will definitely see a difference by investing in a video card. The only thing I still can't figure out is if I should be focusing on getting more VRAM or getting more CUDA cores.
The Tweakers Page explained that if your project was too big and you exhausted the VRAM then it would stop using the GPU for the rest of the render/export. If there was a way I could see how much VRAM my personal projects would require that would be a huge help. Is that related to the export size in any way? Basically, if I looked at how big my export was when it was finished, is that a good gauge to see how much VRAM that project needed?
If I only needed 2 to 3GB then I'd go with the GTX 780 to take advantage of more CUDA cores. If I needed more than 3GB then I'd likely go with the GTX 770 4GB model to avoid the slow down of not being able to use the GPU once it's full.