I know that there are countless topics about this case but that is different. There are mentioned constantly Nvidia 4090 or 4080 but other GPUs asides from the most popular, are mentioned rarely. I think to use professional grade GPU to run TVAI from Quadro and Tesla lines. Especially one particular Tesla model give me impression and that is P100. Here’s the thing. Most GPUs, including modern ones, have Double-Precision Performance in GFLOPS, while Tesla P100 support 4.763 TFLOPS! That is much bigger even when compare to RTX 4090! So my questions are will Topaz run smoothly on a server GPU, will Topaz take an advantege of GPU with such specs, has FP64 performance significant impact in this particular key, will Topaz Video get benefits with TERAflops double precision performance, instead of a couple hundreds Gflops? The drama is not so big with the Quadro cards but I think to get RTX A4000. Is it a good choice? How it is compare with GeForce cards? I know that Quadro cards are more stable and more power efficient but what about performance? And again, is TVAI optimised to work with Quadro cards?
No / yes, depends on age and architecture.
Video AI makes use of Tensors and fp32/fp16 hardware, not FP64 Processing Cores.
Slower.
Its about stability.
Don’t think so much about it.
Quadros are not magic, it’s “only” about stability.
And memory size.
And 24 / 7 / 365 Support.
The P100 is a waste of money, no faster than a 2080.
Had a P5000, 2x Quadro RTX 5000 and a Radeon Pro W6800 in use.
I miss the W6800 because of its memory size.
I use an RTX 4090 and have not regretted buying it, the 5090 is coming in December - January.
Will land at Quadro RTX 5000, 4060ti, performance level.
1 Like
Thank you so much for the quick response and especially for the patience and given information!
1 Like
