GNU Radio or Pothos?


#1

Hello,

I am new to SDR.

Since a few days I am wondering how Pothos compares to GNU Radio.

I understood Pothos has more advanced topology evaluation scheduler and it seems it is more Windows friendly.
It seems it is developed by a very talented and productive person.
From the other hand behind GNU Radio we have big community for years so I guess we will have much bigger functionality there.

Can someone having some experience in those two shed some light on this in terms of performance and functionality (size of the available blocks set)

Thanks
Dimitar


#2

I haven’t had a chance to play with my LimeSDR with GNURadio or Pothos because I’m using SDRConsole V3.0. The only thing I wanted to confirm is that Pothos was authored by a VERY talented fellow who’s tirelessly works on the LimeSDR effort as one of the development team - and that’s @joshblum (Josh Blum). You will not find a more talented and helpful guy than Josh…I know, I’ve worked with him on the LimeSDR Beta effort. If you’re going to use anything, use Pothos…

73 de Marty, KN0CK


#3

Hi Marty,

Yeah I have read about the story behind Pothos and I am already getting some idea and feelings about Josh.
However I want to get some pragmatic thoughts about those two software packages.

Thanks
Dimitar


#4

Great question.
Is anyone doing anything on Pothos in relation to DMR encoding and decoding?
I’ve seen some modules for GNURadio covering that.
Sorry, not trying to hijack the thread, just thought this was a similar topic.


#5

one problem I’m having is that because there are so many abstraction layers between gnuradio and the lms driver, it’s very hard to adjust parameters on the lime that aren’t thought of ahead of time.

For example, in gnuradio the layers look like

gnuradio -> gr-osmosdr -> soapysdr -> libLMS7Support.so -> libLimeSuite.so.17.02-1 (actual driver)

so when we are using a source block for the lime using gr-osmosdr we see the below dialog,
but there is no way to set the TIA Gain.

I think gr-osmosdr (through soapy) technically supports setting any gain, but it’s just not surfaced in the dialog… I’ll have to experiment with modifying the dialog.

So it’s probably possible to modify the python file that gnuradio generates to set the approriate gains/settings.

Of course, it’s isn’t ideal when prototyping, because re-generating the flowgraph then wipes out your modifications.

Note that in Pothos, the SDR Source is setup to read arbitrary JSON in the fields, which helps a little bit:


#6

LimeSDR have two receive channel . I dont know to set device argument in osmocom source for receive independently ? in USRP source IP address distinguishes two receiver but in LimeSDR(USB 3 interface ) i dont know . please help me
Thank


#7

bardia1990, could you please post your question in a separate thread?

Back on the original question. I have already spent few hours with both Gnu Radio and Pothos (mostly on windows, same machine ) and have some “feelings” let say.
(Please don’t ask me to provide arguments)

  • Pothos diagram evaluation and widgets rendering seems to be faster/ more responsive compared with GR
  • Pothos doesn’t have that big block set library
  • Pothos ecosystem development is moving pretty fast (probably faster than GR).

Would you agree/disagree with the above?


#8

disagree. GNUradio has more signal processing block that we can use in all application.


#9

OK, bardia1990, can you please elaborate.