multiple nvidia gpu nvenc

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

multiple nvidia gpu nvenc

GStreamer-devel mailing list
According to this site
https://developer.nvidia.com/video-encode-and-decode-gpu-support-matrix-new

1 nvidia gpu may encode 3 streams at the same time. I have 6 streams and 2
rtx3080, may I know how do I set the environment or pipeline,  so I can
split my processes over two gpus?

thanks



--
Sent from: http://gstreamer-devel.966125.n4.nabble.com/
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: multiple nvidia gpu nvenc

GStreamer-devel mailing list
Not sure, but one thing I would suggest is removing the 3 stream limit. 

https://github.com/keylase/nvidia-patch

><>
Nathan Stratton
Founder, CTO
Vocinity, Inc. 


On Wed, Jun 9, 2021 at 5:20 AM mmssm via gstreamer-devel <[hidden email]> wrote:
According to this site
https://developer.nvidia.com/video-encode-and-decode-gpu-support-matrix-new

1 nvidia gpu may encode 3 streams at the same time. I have 6 streams and 2
rtx3080, may I know how do I set the environment or pipeline,  so I can
split my processes over two gpus?

thanks



--
Sent from: http://gstreamer-devel.966125.n4.nabble.com/
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel

_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: multiple nvidia gpu nvenc

GStreamer-devel mailing list
Thanks for the reply. I dont know the limit can be removed.




I found the loop here

line 130 at
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/blob/master/sys/nvcodec/plugin.c

The loop "for (i = 0; i < dev_count; i++)"  will open a nvenc  session on
all cuda devices for one stream by calling NvEncOpenEncodeSessionEx. Such
concurrent session is limited by nvidia by default.

Is there any reason to request all the cuda devices instead of only one
device?





--
Sent from: http://gstreamer-devel.966125.n4.nabble.com/
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel