Re: [Discuss-gnuradio] fftsink and scopesink refactored

[Martin Dvh]

Jonathan Jacky wrote:
> 
> I tried the new fftsink and scopesink on Mac OS X 10.4.3 Tiger.
> 
> Both of them work without modifications on a 1 GHZ Powerbook G4.
> Previous versions would not always work --- they would start up
> and display a blank window, with the twirling beachball icon that
> indicates the program is busy, apparently because wxPython on the Mac
> couldn't keep up with the data.
> 
> Now scopesink just barely works on the 1 GHZ Powerbook.  With the timebase
> at 100 us/div or more, the display frequently freezes and the beachball
> appears, and the only escape is to type ^C into the terminal
> window.  At smaller values of us/div (fewer cycles on the display) the
> display updates and the controls are responsive.
> 
> scopesink works well (no beachball) on a 1 GHz PowerMac G4 (dual
> processors) and on a 2.7 GHz PowerMac G5 (dual processors).
> 
> Now fftsink works well even on the 1 GHz Powerbook.
> 
> There is one oddity.  When scopesink starts up, the vertical scale is at
> e+35 !  It takes about two minutes to autorange down to
> +/-1000 full scale.  This behavior occurs on both faster and slower Macs.
I have the same problem on linux x86.
Although the problem is not allways there.
I suspect one of the buffers doesn't get initialized at 0.
It seems to also only happen for me when I use the usrp as source with an 
actual signal.
(When I put the usrp in counting mode, the problem goes away and I get a nice 
sawtooth signal)

Greetings,
Martin
> 
> Jon Jacky
> 
> 
> 
> 
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>