TI OMAP DSP support

classic Classic list List threaded Threaded
21 messages Options
12
Reply | Threaded
Open this post in threaded view
|

TI OMAP DSP support

g c-5
I read somewhere that GStreamer has a TI OMAP plug-in
but I couldn't find any mention of it in the overview
of all plug-ins documentation.

1- Can you confirm whether it is available;

If yes:

2a- Does it accelerate video decompression or does it
just streamline some part of the process?

2b- If yes, does it accelerate xvid or ffmpeg
decompression?

3- Does it support the TI OMAP2420?

Thanks in advance for any guidance.


      ___________________________________________________________
Yahoo! Answers - Got a question? Someone out there knows the answer. Try it
now.
http://uk.answers.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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Felipe Contreras
On Jan 16, 2008 6:30 PM, g c <[hidden email]> wrote:
> I read somewhere that GStreamer has a TI OMAP plug-in
> but I couldn't find any mention of it in the overview
> of all plug-ins documentation.

There is different DSP stuff coming from TI. For OMAP they provide
OpenMAX IL wrappers for their DSP codecs, which in turn can be used in
GStreamer with gst-openmax.

> 1- Can you confirm whether it is available;
>
> If yes:

I can't find any information regarding OpenMAX IL support for the
OMAP2420, but it's supposed to be supported:

http://www.khronos.org/consumers/product_details/omap2430_processor/

Perhaps you can ask in customer support.

> 2a- Does it accelerate video decompression or does it
> just streamline some part of the process?

The video decompression is completely executed in DSP side.

> 2b- If yes, does it accelerate xvid or ffmpeg
> decompression?

They'll have to support OpenMAX IL.

> 3- Does it support the TI OMAP2420?
>
> Thanks in advance for any guidance.

I have been using these codecs for a while, but I don't know what's
the public status.

Best regards.

--
Felipe Contreras

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Felipe Contreras
On Jan 16, 2008 8:45 PM, Felipe Contreras <[hidden email]> wrote:

> On Jan 16, 2008 6:30 PM, g c <[hidden email]> wrote:
> > I read somewhere that GStreamer has a TI OMAP plug-in
> > but I couldn't find any mention of it in the overview
> > of all plug-ins documentation.
>
> There is different DSP stuff coming from TI. For OMAP they provide
> OpenMAX IL wrappers for their DSP codecs, which in turn can be used in
> GStreamer with gst-openmax.
>
> > 1- Can you confirm whether it is available;
> >
> > If yes:
>
> I can't find any information regarding OpenMAX IL support for the
> OMAP2420, but it's supposed to be supported:
>
> http://www.khronos.org/consumers/product_details/omap2430_processor/

Found it:
http://www.khronos.org/developers/library/sf_2005/platforms//TI_OMAP-and-Linux.ppt

--
Felipe Contreras

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
In reply to this post by Felipe Contreras

--- Felipe Contreras <[hidden email]>
wrote:
> There is different DSP stuff coming from TI. For
> OMAP they provide
> OpenMAX IL wrappers for their DSP codecs, which in
> turn can be used in
> GStreamer with gst-openmax.

In the following article TI says they have a plug-in
for GStreamer:
[hidden email]

> They'll have to support OpenMAX IL.

So if I understand what you are saying, it is that
Gstreamer currently does not have any OMAP DSP
accelerated video decode capability. That would have
to be supported by a specific codec and that codec
would have to be available as a plug-in to GStreamer.

That's what I would expect. I couldn't imagine how one
GStreamer plug-in could provide video decompression
acceleration for another plug-in and yet that's what
the above article seems to state.

However, the Nokia N800 and N810 must surely have a
media player that uses the acceleration capability of
the OMAP 2420 chip.

Regards

Graham


      __________________________________________________________
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Felipe Contreras
On Jan 16, 2008 9:22 PM, g <[hidden email]> wrote:

>
> --- Felipe Contreras <[hidden email]>
> wrote:
> > There is different DSP stuff coming from TI. For
> > OMAP they provide
> > OpenMAX IL wrappers for their DSP codecs, which in
> > turn can be used in
> > GStreamer with gst-openmax.
>
> In the following article TI says they have a plug-in
> for GStreamer:
> [hidden email]

Wrong link?

> > They'll have to support OpenMAX IL.
>
> So if I understand what you are saying, it is that
> Gstreamer currently does not have any OMAP DSP
> accelerated video decode capability. That would have
> to be supported by a specific codec and that codec
> would have to be available as a plug-in to GStreamer.
>
> That's what I would expect. I couldn't imagine how one
> GStreamer plug-in could provide video decompression
> acceleration for another plug-in and yet that's what
> the above article seems to state.

Er, no idea, would have to read the article.

> However, the Nokia N800 and N810 must surely have a
> media player that uses the acceleration capability of
> the OMAP 2420 chip.

For video? Yeah that would be great, but that would require either:

a) Someone to develop the DSP codecs for the dsp-gw [1] and ARM side wrappers.

b) Nokia to provide the TI DSP stack (dsp codecs, arm/dsp-bridge,
openmax wrappers), and support for the N8x0 (gst-openmax).

c) TI to provide their DSP stack publicly and some maemo hackers
provide support for the N8x0.

I can't comment on what has more chances to happen, but it's always
good to shout to the companies for answers.

[1] http://dspgateway.sourceforge.net/pub/index.php

--
Felipe Contreras

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Stefan Sauer
In reply to this post by g c-5
hi,

g schrieb:

> --- Felipe Contreras <[hidden email]>
> wrote:
>> There is different DSP stuff coming from TI. For
>> OMAP they provide
>> OpenMAX IL wrappers for their DSP codecs, which in
>> turn can be used in
>> GStreamer with gst-openmax.
>
> In the following article TI says they have a plug-in
> for GStreamer:
> [hidden email]
>
>> They'll have to support OpenMAX IL.
>
> So if I understand what you are saying, it is that
> Gstreamer currently does not have any OMAP DSP
> accelerated video decode capability. That would have
> to be supported by a specific codec and that codec
> would have to be available as a plug-in to GStreamer.
>
> That's what I would expect. I couldn't imagine how one
> GStreamer plug-in could provide video decompression
> acceleration for another plug-in and yet that's what
> the above article seems to state.
>
> However, the Nokia N800 and N810 must surely have a
> media player that uses the acceleration capability of
> the OMAP 2420 chip.

the video decoding for OS2008 (N8x0) uses arm codecs. The voip codecs currently
run on th dsp (together with the rest of the audio subsystem).

Stefan

>
> Regards
>
> Graham
>
>
>       __________________________________________________________
> 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
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel


-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
In reply to this post by Felipe Contreras
--- Felipe Contreras <[hidden email]>
wrote:
> > In the following article TI says they have a
> plug-in
> > for GStreamer:
> > [hidden email]
>
> Wrong link?

Duh. Sorry.
http://www.eetimes.com/news/design/showArticle.jhtml;jsessionid=NPKIHIAW3OOCOQSNDLRSKH0CJUNN2JVN?articleID=193401461&pgno=1

>
> For video? Yeah that would be great, but that would
> require either:
>
> a) Someone to develop the DSP codecs for the dsp-gw
> [1] and ARM side wrappers.
>
> b) Nokia to provide the TI DSP stack (dsp codecs,
> arm/dsp-bridge,
> openmax wrappers), and support for the N8x0
> (gst-openmax).
>
> c) TI to provide their DSP stack publicly and some
> maemo hackers
> provide support for the N8x0.

TI provides a full SDK:
http://focus.ti.com/general/docs/wtbu/wtbuproductcontent.tsp?contentId=4671&navigationId=11990&templateId=6123

I would have thought that Nokia would have provided a
media player that supported their own DSP. Without it,
media playback would be rather poor.

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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
In reply to this post by Stefan Sauer
--- Stefan Kost <[hidden email]> wrote:
> the video decoding for OS2008 (N8x0) uses arm
> codecs. The voip codecs currently
> run on th dsp (together with the rest of the audio
> subsystem).

Thanks for the info.

So how come there is no DSP assistance for decoding?

When you say that OS2008 uses ARM codecs, does that
mean that I should use xvid or ffmpeg as is in
GStreamer with no modification?

Can you tell me what performance to expect? I have
previously been told that even the N810 can only
support VGA at 20 fps even though TI says their OMAP
2420 can support 30fps.

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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Stefan Sauer
hi,

g schrieb:
> --- Stefan Kost <[hidden email]> wrote:
>> the video decoding for OS2008 (N8x0) uses arm
>> codecs. The voip codecs currently
>> run on th dsp (together with the rest of the audio
>> subsystem).
>
> Thanks for the info.
>
> So how come there is no DSP assistance for decoding?

technically it would be good to have video codecs on the dsp. if this is not
offered at the time its needed, one has to look for alternatives. most of this
is business related and you will find out about that when your trying it :(

>
> When you say that OS2008 uses ARM codecs, does that
> mean that I should use xvid or ffmpeg as is in
> GStreamer with no modification?
>
That would be a bit naive. The amr11 that is on the TI OMAP2420 has vector
operations and you can only get good performance if you can make use of them.

>
> Can you tell me what performance to expect? I have
> previously been told that even the N810 can only
> support VGA at 20 fps even though TI says their OMAP
> 2420 can support 30fps.

TI says many things. They need to sell their chips,

Stefan

>
> 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
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel


-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
> That would be a bit naive. The amr11 that is on the
> TI OMAP2420 has vector
> operations and you can only get good performance if
> you can make use of them.

ffmpeg and xvid are available as plug-ins for
GStreamer and presumably work on Maemo 2008. What I
meant is that they presumably use ARM instructions.
The point is whether they ought to be modified to make
use of the DSP part of OMAP2420 to improve
performance.

> >
> > Can you tell me what performance to expect? I have
> > previously been told that even the N810 can only
> > support VGA at 20 fps even though TI says their
> OMAP
> > 2420 can support 30fps.
>
> TI says many things. They need to sell their chips,

Again, the point is the difference in performance with
/ without the DSP part of OMAP 2420. TI says that if
you use the DSP part you can get 30fps. If the current
implementations of media players do not use the DSP
part and only ARM, what is the performance?

I would be surprised if there is no software that uses
such a powerful processing element for media playback.

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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Josep Torra Valles
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.

> > That would be a bit naive. The amr11 that is on the
> > TI OMAP2420 has vector
> > operations and you can only get good performance if
> > you can make use of them.
>
> ffmpeg and xvid are available as plug-ins for
> GStreamer and presumably work on Maemo 2008. What I
> meant is that they presumably use ARM instructions.
> The point is whether they ought to be modified to make
> use of the DSP part of OMAP2420 to improve
> performance.

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.
The ARM6 core have a set of media instructions that can be used to
accelerate the decoding process, it's like mmx on x86, you need to write
assembly code in order to take advantage of it.  

> > >
> > > Can you tell me what performance to expect? I have
> > > previously been told that even the N810 can only
> > > support VGA at 20 fps even though TI says their
> > OMAP
> > > 2420 can support 30fps.
> >
> > TI says many things. They need to sell their chips,
>
> Again, the point is the difference in performance with
> / without the DSP part of OMAP 2420. TI says that if
> you use the DSP part you can get 30fps. If the current
> implementations of media players do not use the DSP
> part and only ARM, what is the performance?
>
> I would be surprised if there is no software that uses
> such a powerful processing element for media playback.

The OMAP solution provide an LCD controller/video frame buffer but
unfortunately it can't drive the screen of 800x480 pixels used on the
N8x0, to solve this an external controller had to be added to the
device.

In the TI article it's considered 0 memcpy and mem mapping in the video
framebuffer for the ARM and DSP cores, this is not possible in the nokia
device, 1 memcpy needs to be done to the external framebuffer and this
will limit (bus speed) the performance and in consequence the  video
size and fps that can be played on the device.

I've been reading the capabilities of the DSP some time ago and aren't
as impressive as you think.

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).




-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
--- Josep Torra Valles <[hidden email]> 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.

> > 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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Stefan Sauer
hi,

Quoting g <[hidden email]>:

> --- Josep Torra Valles <[hidden email]> 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.

Stefan

>
>> > 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
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
>



-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Felipe Contreras
In reply to this post by g c-5
On Jan 17, 2008 1:42 AM, g <[hidden email]> wrote:

> --- Josep Torra Valles <[hidden email]> 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.
>
> > > 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.

That article is talking about the DaVinci platform, which is not using OMAP2420.

For DaVinci AFAIK TI developed it's own codec API, something like
OpenMAX IL, and then developed their own GStreamer elements, like
gst-openmax.

--
Felipe Contreras

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Kalle Vahlman
In reply to this post by g c-5
2008/1/17, g <[hidden email]>:
> 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.

Anyone can write a GStreamer plugin to do whatever task they need to,
it doesn't mean it will be available from the GStreamer site.

> --- Josep Torra Valles <[hidden email]> wrote:
> > 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.

What is perhaps not obvious is that whatever TI promises for it's SOC
isn't necessarilary true for any given product that uses the SOC.

For example, if talking about the N8x0 platforms, they have
limitations to their implementation (both hw and sw) that in short
mean:

 - As mentioned, the external LCD controller has lousy memory
bandwidth which limits the framerate even if you could decode faster
 - If a "dynamic" DSP task is loaded, the ARM core is locked to 333
MHz (max is 400MHz)

This in effect means that the usage of DSP for MP3 decoding "wastes"
77MHz of the ARM core which, according to a mplayer dev, is more than
enough to decode MP3 audio. Thus, you can achieve better video
performance by using a software codec to decode both the video and the
audio than what you get by using the DSP codec for the MP3 decoding.
Search the maemo-developers mailing list archives for more details.

I don't know if a DSP video codec would change that scenario though,
but for N8x0 that point is pretty moot since the actual bottleneck is
the slow memory bus to the LCD controller, not decoding power (as far
as I know at least).

Oh, and for the record, despite all the shortcomings I'm happy enough
watching 640x240@23.976fps videos (scaled to whatever the aspect ratio
really is) on my N800 :)

--
Kalle Vahlman, [hidden email]
Powered by http://movial.fi
Interesting stuff at http://syslog.movial.fi

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
In reply to this post by Stefan Sauer
Stefan, Kalle, Felipe

Thank you all for your help.

So it is as I expected re. DSP support in GStreamer.
Someone in another group had said that GStreamer
supported the DSP for the N810 and then I found the TI
article on the GStreamer site so I was seeking
confirmation.

The information about bandwidth to the screen is
useful.

I think the N810 is a bit more powerful than the N800.
I guess the next steps for me will be to write a
simple test that simply refreshes the screen as
quickly as possible to confirm the screen bottleneck.
Next will be to try and implement iDCT on the DSP.

g.

--- Stefan Kost <[hidden email]> wrote:

> hi,
>
> Quoting g <[hidden email]>:
> > --- Josep Torra Valles <[hidden email]>
> 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.
>
> Stefan
> >
> >> > 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
> > [hidden email]
> >
>
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
> >
>
>
>
>
-------------------------------------------------------------------------
> 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
> [hidden email]
>
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
>



      __________________________________________________________
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
In reply to this post by g c-5
> Stefan, Kalle, Felipe
>
> Thank you all for your help.
>
> So it is as I expected re. DSP support in GStreamer.
> Someone in another group had said that GStreamer
> supported the DSP for the N810 and then I found the
> TI
> article on the GStreamer site so I was seeking
> confirmation.
>
Okay, I'm confused again. In the Maemo Quick Start
Guide, it states, "Maemo has a high level programming
support for the DSP in form of GStreamer elements
which can decode most of the supported file formats.
By using the DSP computing load is also taken off from
the main processor which greatly enhances system
performance and responsivity."

So which are those elements?

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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Felipe Contreras
On Jan 18, 2008 4:08 PM, g <[hidden email]> wrote:

> > Stefan, Kalle, Felipe
> >
> > Thank you all for your help.
> >
> > So it is as I expected re. DSP support in GStreamer.
> > Someone in another group had said that GStreamer
> > supported the DSP for the N810 and then I found the
> > TI
> > article on the GStreamer site so I was seeking
> > confirmation.
> >
> Okay, I'm confused again. In the Maemo Quick Start
> Guide, it states, "Maemo has a high level programming
> support for the DSP in form of GStreamer elements
> which can decode most of the supported file formats.
> By using the DSP computing load is also taken off from
> the main processor which greatly enhances system
> performance and responsivity."
>
> So which are those elements?

http://maemo.org/development/documentation/how-tos/3-x/multimedia_architecture.html

--
Felipe Contreras

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

g c-5
> > By using the DSP computing load is also taken off
> from
> > the main processor which greatly enhances system
> > performance and responsivity."
> >
> > So which are those elements?
>
>
http://maemo.org/development/documentation/how-tos/3-x/multimedia_architecture.html

Thanks but according to that diagram the plug-in
hantro4100dec handles MPEG4 video decoding. However,
it is not in the list plug-ins for GStreamer, nor did
it come up when I Googled the entire GStreamer site.

http://gstreamer.freedesktop.org/documentation/plugins.html

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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
Reply | Threaded
Open this post in threaded view
|

Re: TI OMAP DSP support

Felipe Contreras
On Jan 18, 2008 5:23 PM, g <[hidden email]> wrote:

> > > By using the DSP computing load is also taken off
> > from
> > > the main processor which greatly enhances system
> > > performance and responsivity."
> > >
> > > So which are those elements?
> >
> >
> http://maemo.org/development/documentation/how-tos/3-x/multimedia_architecture.html
>
> Thanks but according to that diagram the plug-in
> hantro4100dec handles MPEG4 video decoding. However,
> it is not in the list plug-ins for GStreamer, nor did
> it come up when I Googled the entire GStreamer site.
>
> http://gstreamer.freedesktop.org/documentation/plugins.html

Because it's not provided by the GStreamer community nor it's open,
it's closed source provided by hantro.

--
Felipe Contreras

-------------------------------------------------------------------------
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
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
12