The GPU is used by very few things in After Effects. Unless you are using one very specific feature (the ray-traced 3D renderer), the additional GPU is not going to give you better performance in After Effects.
See this page for details:
On that page, you'll see this about dual GPUs:
"If you have multiple GPUs installed, the GPU-accelerated ray-traced 3D renderer will use the CUDA cores on all of them, as long as they are of the same CUDA compute level. (See the technical specifications of your GPU for its CUDA compute level.) After Effects will also use all of the VRAM on the installed GPUs, with the caveat that both cards will be treated as if they each have the amount of VRAM on the card with the lesser amount of VRAM."
I may have framed that question wrong. Yes, I'm going to do a TON of 3D rendering with both the built in raytracer and C4D. The question is, I haven't seen any proof that all the CUDA processors are being used, or even being reported as available. I use other 3D software which reports OpenCL profiles and when I add the extra card it shows no additional interaction or performance enhancements. So I guess what I should be asking is there a third party profiler that can report if AE and other programs are taking advantage of SLI. Some of the newer card are supporting 3-way SLI but what difference does it make if the software is not designed to use it?
1 person found this helpful
I am not sure that I know what you are looking for but if you get a copy of GPU-Z and run an instance of it for both GPU's you can see the usage of each one. Here is a sample of my laptop GTX 765M running at 99% GPU load.
But those 8800 GTS antiques are pretty puny and will not work to provide any Premiere MPE acceleration. They are not CUDA equipped and not enough RAM and forget SLI for CUDA work.
Yes, thanks! That's something I'll look into. The 8800 do have CUDA processor and are SLI ready, yet they are small in comparison. Total CUDA's 384, with 384 bit bandwidth. They actually work pretty good for their age and specs, as well were dirt cheap. I bought them to test out network rendering of some other applications when I noticed that nothing seemed to take advantage of the SLI, so I assumed if the software wasn't written to work with it then you are S.O.L. I am not a gamer and have nothing that is known to work with multiple graphic cards as a benchmark. This application looks like it would shed some
light regarding preformance.
I just would like to see proof positive that all this technology is working as advertised. Could you imagine the costly mistake of investing in multiple high-end cards only to find connecting them with SLI makes no difference with AE?
1 person found this helpful
Maybe you're still not asking exactly what you are wanting to know!?
Creative Cloud AE does indeed tap multiple GPUs. My testing was done on a fast, dual-Xeon PC with dual GTX Titan video cards. I did not configure with SLI anything - no bridge cable and no SLI configuration with the nVidia drivers, etc.
I used a AE Raytrace benchmark for my "playing". You may want to check out: Teddy Gage AE benchmark.
Premiere Pro also taps the power of multiple GPUs for renders without SLI enabled. What you are rendering makes a huge difference whether or not this is worthwhile though; GPUs help massively doing timeline outputs to DVD-MPEG whereas the CPU capabilities are what your need to render timeline outputs to Blu-ray. BTW, prior to the CC versions, Premiere Pro could only tap a single GPU so your testing must be with the newest Adobe software if you want dual-GPU testing to be meaningful.
Regarding how to monitor how hard your CPU and GPU are working, you need to use built-in Windows monitoring tools and/or some of the 3rd party monitoriing tools.
Wow...that is a very informative read. I wish I had waited, just bought a gtx780. I see I'm going to have issues with the CPU and Ram now. I guess this will be a "one hurdle at a time" resolution until I have a highly productive machine.
SLI is completely different than GPU acceleration. You dont require the cards setup in SLI to accelerate accross more than 1 card nor does it add that feature. The application has to be coded via CUDA or Open CL to scale accross the GPU's. That is the same if you use Octane as well for C4D which I would suggest you look into if you really want GPU acceleration with C4D.