Dual 6th Gen NVENC (Pascal) encoding speed

aberniando

New Member
We all know that 7th Gen NVENC introduced on Turing was amazing in terms of quality.

But I just found out that some Pascal 1070/1080/Ti are all armed with TWO NVENC chip.

Makes me wonder how FAST those 2 6th Gen Pascal NVENC chip compared to 7th Gen NVENC Turing chip (all Turing and Ampere only have 1 NVENC).

I hope one of you guys here have ever give it a try. thanks!
 

DayGeckoArt

Member
It might only affect the number of streams you can record because the actual processing is done by the CUDA cores
 

aberniando

New Member
It might only affect the number of streams you can record because the actual processing is done by the CUDA cores
Screenshot_20230106-063920_Chrome.jpg

dont think so my friend, nvenc is independent from graphic
 

sandrix

Member

Encoder Features using CUDA

Although the core video encoder hardware on GPU is completely independent of CUDA cores or graphics engine on the GPU, following encoder features internally use CUDA for hardware acceleration.
Note: The impact of enabling these features on overall CUDA or graphics performance is minimal, and this list is provided purely for information purposes.
  • Two-pass rate control modes for high quality presets
  • Look-ahead
  • All adaptive quantization modes
  • Weighted prediction
  • Encoding of RGB contents
 

DayGeckoArt

Member

Encoder Features using CUDA

Although the core video encoder hardware on GPU is completely independent of CUDA cores or graphics engine on the GPU, following encoder features internally use CUDA for hardware acceleration.
Note: The impact of enabling these features on overall CUDA or graphics performance is minimal, and this list is provided purely for information purposes.
  • Two-pass rate control modes for high quality presets
  • Look-ahead
  • All adaptive quantization modes
  • Weighted prediction
  • Encoding of RGB contents

Thanks, that's helpful because people always cite Nvidia's documents that claim Nvenc is totally separate.

Even here, they say it's "completely independent" then go into how it's not completely independent!

In the screenshot above I showed 7% CUDA usage on an RTX 2080 Ti, but that's just recording a static web browser. In games, I consistently see about 50 watts additional power usage and higher CUDA %, so I also think it's interesting they say "minimal" which is a subjective word.
 
Top