There is some discussion of the cache maintenance overhead inherent in
multi-processor systems in the archives here:
http://linux.omap.com/pipermail/davinci-linux-open-source/2007-May/00327
0.html
There are some techniques to eliminate some of this cache maintenance if
possible, ranging from simply understanding what's going on WRT cache
management and tuning your algorithm appropriately:
http://wiki.davincidsp.com/index.php?title=Cache_Management
... to hacking the cache management completely out of the system:
http://wiki.davincidsp.com/index.php?title=Overriding_stubs_and_skeleton
s
Note that removing necessary cache management may result in, at best,
incorrect behavior and potentially system instability.
Chris
________________________________
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On
Behalf Of ilay nissim
Sent: Tuesday, December 25, 2007 2:23 AM
To: [email protected]
Cc: [email protected]
Subject: DM6446 - ARM TO DSP communication overhead using XDM
Hi All,
I have developed an xdm component.
when i run it as an empty component, meaning in the dsp i
immediately return
i still see a 2 ms overhead.
I am sending a video frame with the size of 720X480.
My time measurements are on the arm.
pseudo code:
get time()
VIDENC_PROCESS
get time()
Thank you.
Ilay
________________________________
Be a better friend, newshound, and know-it-all with Yahoo!
Mobile. Try it now.
<http://us.rd.yahoo.com/evt=51733/*http://mobile.yahoo.com/;_ylt=Ahu06i6
2sR8HDtDypao8Wcj9tAcJ>
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
