Mailinglist Archive: opensuse (908 mails)

< Previous Next >
Re: [opensuse] Graphics for Server
On 2016-06-01 07:32, Per Jessen wrote:
Per Jessen wrote:

Carlos E. R. wrote:


The graphics card is not used at all when converting videos to
another format. Only when displaying that video.

A graphics card's GPUs might well be used during
encoding/decoding/conversion. For instance, ffmpeg is able to use
vaapi.

Paul, your systems are virtual identical. To test if it is an issue
with the graphics card, it might be worth swapping the graphics cards.
Also try "vainfo" to see if video hardware acceleration is in
use/available.

Uh, with Nvidia cards, it's called VDPAU instead of VAAPI, I don't know
if vainfo is aware or if there is a dedicated utility for vdpau.

I don't even have vainfo installed. :-?

vainfo comes with "vaapi-tools" which is really for Intel Graphics only.
I'm sure there must be something similar for Nvidia.

I have no idea about that. There is a "vdpauinfo" here. It prints a page
of information, but nothing about ffmpeg using it or not. :-?

"man ffmpeg" has nothing about nvidia. There is a mention about vdpau,
though:

-hwaccel[:stream_specifier] hwaccel
(input,per-stream)
Use hardware acceleration to decode the
matching stream(s). The allowed values of
hwaccel are:

none
Do not use any hardware acceleration
(the default).

auto
Automatically select the hardware
acceleration method.

vda Use Apple VDA hardware acceleration.

vdpau
Use VDPAU (Video Decode and
Presentation API for Unix) hardware
acceleration.

dxva2
Use DXVA2 (DirectX Video Acceleration)
hardware acceleration.

qsv Use the Intel QuickSync Video
acceleration for video transcoding.

Unlike most other values, this option
does not enable accelerated decoding
(that is used automatically whenever a
qsv decoder is selected), but
accelerated transcoding, without
copying the frames into the system
memory.

For it to work, both the decoder and
the encoder must support QSV
acceleration and no filters must be
used.

This option has no effect if the selected
hwaccel is not available or not supported
by the chosen decoder.

Note that most acceleration methods are
intended for playback and will not be
faster than software decoding on modern
CPUs. Additionally, ffmpeg will usually
need to copy the decoded frames from the
GPU memory into the system memory,
resulting in further performance loss. This
option is thus mainly useful for testing.



Note that it is not enabled by default, and only used for playback, and
that it says: "will not be faster than software decoding on modern CPUs"


-hwaccel_device[:stream_specifier]
hwaccel_device (input,per-stream)
Select a device to use for hardware
acceleration.

...

-hwaccels
List all hardware acceleration methods
supported in this build of ffmpeg.


The speed of ffmpeg varies a lot depending on codec selected and
options. Some codecs parallelize wonderfully. Others very badly, and for
those it is better to convert several media streams at the same time,
one per cpu core. If the CPU is fast enough, it is the hard disk which
becomes a limitation.


--
Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 "Bottle" at Telcontar)

< Previous Next >
List Navigation
Follow Ups