[gst-devel] TI OMAP DSP support

Stefan Kost ensonic at hora-obscura.de
Thu Jan 17 08:57:34 CET 2008


Quoting g <the_ether2 at yahoo.co.uk>:
> --- Josep Torra Valles <j.torra at telefonica.net> wrote:
>> First I think that you should get more knowledge
>> about the N8x0
>> architecture and the OMAP SoC in order to be able to
>> understand the
>> scenario for your project.
> My "project" is simply to decode MPEG video. As stated
> previously the core elements I intend to use are
> already available as plug-ins to GStreamer.
> My questions have been whether GStreamer already has
> implemented any DSP acceleration and if not, what the
> performance hit of their absence is compared to the
> potential stated by TI.
GStreamer itself does not implement any for of hardware specific acceleration.
A few elements use liboil which helps to benefit from usinf vector  
operations of CPUs (MMX, SSE2, ...)

If you want to benefit from a DSP, you have to rewrite the codecs  
using DSP nstructions. You cannot just recompile it for a DSP and you  
cannot just runs some parts easily on a DSP. So nothing Gstreamer can  
do here. You have to get codec implementations for the DSP in question  
by yourself. Thats why using a DSP is so difficult and often projects  
decide to better use the CPU and invest in good compilers.

>> > ffmpeg and xvid are available as plug-ins for
>> > GStreamer and presumably work on Maemo 2008. What
>> If just crosscompile you will have something
>> working, but it doesn't
>> mean that the performance will be enough to get VGA
>> at 30 fps.
> As stated above, the MPEG decoders ffmpeg and xvid are
> already available as plug-ins for GStreamer. I have
> merely asked what the performance is. Maybe someone
> has already optimised one or both of these for ARM and
> / or the DSP portion of the OMAP2420.
> As previously mentioned, a TI article specifically
> states that there is a GStreamer plug-in that uses the
> DSP and yet I cannot find any mention of its existance
> on the GStreamer site.
>> I also wrote some assembly to accelerate video
>> decoding on ARM5e and
>> ARM6 cores and I think that probably on ARM6(400
>> MHz) you can achieve
>> more video decoding speed than using the DSP(133
>> MHz).
> The clock frequency isn't the only part. What matters
> is what work is done per cycle. The DSP obviously has
> an important part to play. That's why it's there in
> the first place. Given the choice, doing all the
> processing purely on the ARM core and not using the
> DSP for any task is silly - especially for video
> decoding.
> g.
>       __________________________________________________________
> Sent from Yahoo! Mail - a smarter inbox http://uk.mail.yahoo.com
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> gstreamer-devel mailing list
> gstreamer-devel at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

More information about the gstreamer-devel mailing list