Questions about certain nVidia GPUs

[Toxikaraidur]

I have a build planned for later down the road involving heavy computational loads such as light/heavy threaded programming, 3D modelling, animation, CAD, and video rendering. Basically, I will be attempting to make some games and other graphical programs. When it comes to playing games, the 1080 Ti is *nearly* undisputed king. but it only looks that fantastic with common FP16 performance which most gaming is. Rendering, modelling, and animation from what I understand are much more compute intensive and take much longer. So while playing games is great, looking at the 1080 Ti and even Titan (whatever the current version is, and no I don't mean the Titan V) seem to fall short when you look at this thread.

https://www.microway.com/knowledge-center-articles/comparison-of-nvidia-geforce-gpus-and-nvidia-tesla-gpus/

less than a TFlop of compute performance on anything less than FP16; I don't plan on waiting around for my work to compile/render/load for a decade, so I'm wondering about two other GPU's on that list.

The NVIDIA Quadro GP100
The NVIDIA Tesla V100

Very similarly specced for two different branches of pro/ai gpus. They would more than double, nearly triple the cost of the rig I plan, but in the long run with aforementioned workloads, would it work?

I know it's a silly question if you consider it, but It is a question I would like answered. Thank you all.

 

[renz496]

probably you miss something? just because GP100 and GV100 is more "compute" oriented means it will will work better in all compute related work. take FP32 for example. Titan Xp have more FP32 performance than GP100. so if your work load only take advantage of FP32 titan Xp still better in that regard (this still not counting driver optimization though).