Hi all,
Thanks to all who responded about the small footprint filesystem; they were all
very helpful. I have a new question:
I want to improve the performance of the CE Server on the DSP side; I know
there are L1 and L2 caches and the
actual amount can be configured in the *.tcf file of the CE server.
1) Do I have to activate these caches in my DSP code or are they already set up
to run automatically by the CE
Server initialization routine?
2) Also, I may want to use some of the L2 cache as IRAM or some of the L1 cache
as L1SDRAM (I know how to
configure these in the *.tcf file) but I don't know how to declare my data to
map to these fast memory regions.
For example, let's say my encoder codec has an array and I want fast access to
it, it is declared as follows:
XDAS_Int32 myArray[100]; // 400 bytes
How do create a link file to force "myArray" to link to the beginning L1SRAM?
The DVEVM codecs / servers
examples automatically generate all the link files (*.cmd) in package/cfg, so
I'm assuming it has to be done
somewhere in one of the *.bld, *.cfg, or *.tcf files, am I correct? Can anyone
please give me some pointers on how to
link data to IRAM or L1SDRAM?
Thanks!
Regards,
Andy_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
