[hpsdr] O/J Power SDR CPU usage

[Bill Tracey]

Hmmm ... not sure I follow.   I'm pretty sure the code does not spend 
a lot of cycles multiplexing/demultiplexing the control data stream 
from the audio data stream -- profiling would  tell for sure.

My guess on the cause of the increased CPU usage is the small fame 
size we're using on the USB bus  -- currently 512 bytes.   We might 
get a perf boost by going to a larger frame size - the current frame 
size goes back to the Xylo days when we had issues writing/reading 
anything much bigger than that reliably.  We've learned a lot since 
then and could prob get a bit of a perf boost with larger frame size.

Regards,

Bill (kd5tfd)

At 05:03 PM 6/25/2007, Philip Covington wrote:
>***** High Performance Software Defined Radio Discussion List *****
>  <....snip....>
>
>I believe this is the big downside to mixing control data in with the
>datastream.
>
>Phil N8VB



 1182813089.0