Posts: 656
Threads: 68
Joined: Feb 2020
Thanks for the new dev version.
I opened an issue to Rigaya regarding the filter usage:
https://github.com/rigaya/NVEnc/issues/592
Dan
Posts: 656
Threads: 68
Joined: Feb 2020
Hello Selur,
after having opened the issue on NVEncC I discovered the there is the option "--vpy-mt" which improve the speed significantly.
Using a simple script with the following code
# loading plugins
core.std.LoadPlugin(path="D:/Programs/Hybrid/64bit/vsfilters/Support/fmtconv.dll")
# Resizing using 10 - bicubic spline
clip = core.fmtc.resample(clip=clip, kernel="spline16", w=1280, h=900, interlaced=False, interlacedd=False)
I obtained the following results
As you can see with parameter "--vpy-mt" nvencc is about 40% faster than vspipe.
Could you add support to "--vpy-mt" ?
Thanks,
Dan
Posts: 10.522
Threads: 56
Joined: May 2017
28.06.2024, 09:40
(This post was last modified: 28.06.2024, 09:42 by Selur.)
I'll look at it after work.
![Smile Smile](https://forum.selur.net/images/smilies/smile.png)
(--vpy-mt isn't mentioned in
https://github.com/rigaya/NVEnc/blob/mas...ions.en.md)
Posts: 656
Threads: 68
Joined: Feb 2020
(28.06.2024, 09:40)Selur Wrote: I'll look at it after work. ![Smile Smile](https://forum.selur.net/images/smilies/smile.png)
(--vpy-mt isn't mentioned in https://github.com/rigaya/NVEnc/blob/mas...ions.en.md)
Rigaya forgot to add it in the documentation, see:
https://github.com/rigaya/NVEnc/issues/591
Dan
P.S.
The new dev version with "--vpy-mt" support is working.
EDIT: Rigaya said that she forgot to remove from the documentation of QSVEnc, she is not sure the "--vpy-mt" is always working.