Re: [Discuss-gnuradio] Problems running example audio_fft.py

[Dave hartzell]

Michael,

Comments/questions below:

On 10/23/06, Michael Dickens <address@hidden> wrote:
> Hence I would believe that if you're running on a single-cpu G4 or
> G5, even up to 2 GHz, the wxGUI stuff probably won't be very
> effective.  OTOH, running a dual-G4 @ 1.25 GHz -just barely- works,
> so running a Dual-G5 @ 2.0 GHz, or the new Core-Duo's at 2.0 GHz,
> will surely be good enough to get a reasonable display - but nothing
> special compared do the same hardware running Linux.

OK, that is interesting....are these things (G4s) really that slow?
Or is something killing the performance in the middle on a Mac?

I know how to limit FFT CPU utilization via the window sizes, but is
there a way to limit the WxPython window updates to something lower?

> I would be quite curious about the CPU utilization of a Linux box -
> along with the hardware specs to make the judgement relatively
> correct.  Even better would be to try an Intel Mac running both OSX
> and Linux and compare the CPU utilization ... hmmm ... that would be
> pretty cool so I'll look into it. - MLD

I just posted another message on this last night (haven't seen it come
across the list yet) but my Fry's budget linux box (AMD 2200+ 1.5 GHz
no-frills w/onboard video) can't quite handle the usrp_wfm_rcv.py
script.  I get a slight pause in the FFT window update, a slight pause
in the audio playback and underruns/overruns (OouOou) on the CLI.  The
CPU hits about 65% with the "python" process and Xorg takes the rest
of the CPU (35%) according to Top.   The usrp_wfm_nogui puts out audio
fine, at about 35% cpu load if I recall.

It seems that the WxPython is quite a resource consumer on any platform...

Dave