Re: [Discuss-gnuradio] oprofile inband code results

[George Nychis]

Eric,

> Before even thinking about libpmt, it would be wise to figure out the
> much greater CPU consumption in libc.  Start with the big offenders.
>
> When that's sorted out (and we're confident about your profile
> numbers) we can have a discussion about what's next.

Sounds good.

> BTW, what is the test case that you're running?
> What is the exact sequence of commands that you are issuing to
> generate this trace?  I'd like to try to reproduce this on my machines.

So actually I was using reset, but said dump.

Heres my exact sequence:

sudo opcontrol --start
sudo opcontrol --reset
./test_usrp_inband_underrun (interp/decim @ 16, 30e6 samples)
sudo opcontrol --save=inband_tx_rx_5
cd .libs/
opreport -rl session:inband_tx_rx_5 lt-test_usrp_inband_underrun &> \
   inband_tx_rx_5

If i run another test, I omit the start :P

And, heres the results of doing just that:
http://www.andrew.cmu.edu/user/gnychis/inband_tx_rx_5

Heres the rbf I'm using, you can change it by editing the underrun 
application, the current checked in copy of my branch is using 
nanocell9, just switch it.
http://www.andrew.cmu.edu/user/gnychis/cs1.rbf

A new run with test_usrp_inband_tx:
http://www.andrew.cmu.edu/user/gnychis/inband_tx_6

The key difference between the two is underrun has both TX and RX 
running, TX only has TX running... and also underrun generates one large 
message of size d_nsamples_to_send, where the normal TX program breaks 
d_nsamples_to_send down into multiple frames.

- George