[hpsdr] NVIDIA Jetson Nano SBC - YouTube video

[ad0es]

Hi,

I wrote a suite of programs running on large hardware/linux. 
Specifically an Asus X99 Server class motherboard
with Nvidia gtx1080 GPU board which has 2560 cores.  I was up to approx. 
55 DDCs (ie. receivers).
I used the ghpSDR3-Alex model for putting together a variety of clients. 
among these were fldigi, Thetis, wsjtx,
GQrx, and GNU Radio Companion.

At the point of 55 DDCs the process was near the end of the second, ie 
it would become "less than realtime" if running more.
Also note, that I am referring to 55 IQ data streams, NOT 55 clients,  
The desktop doesn't have enough power or
screen real estate to run 55 copies of fldigi or GQrx.  You would have 
to distribute the clients among multiple desktops
for that (and ghpSDR3 can distribute the data to other desktops quite 
nicely).

The modified ghpSDR3 approach allows you to use any of the old-school 
programs that process PCM audio from the
sound stack, as well as programs that can directly consume IQ data.

Specifically I used GNURC to build  wide bandwidth panadapters. Since 
GNURC has filters to directly process IQ
data its a good fit.  The biggest drawback to GNURC is that it doesn't 
yet use CUDA for its processing, and thus is
fairly slow.  I got beautiful, fast panadapters using GQrx.

Steve ad0es


On 04/29/2019 07:59 AM, Joe Martin wrote:
> The problem in utilizing GNURadio for typical ham use is twofold as I see it though:
>
>
> Just thought I’d mention it, not that I think the GNURadio approach is viable for everyone.  It isn’t.  But it’s incredibly powerful and doesn’t take a lot of effort to implement basic functionalities.  It is certainly one example of how multiple cores are currently being utilized though, which is to the point of your question I think.
>
> 73, Joe K5SO
>
>
>
>> On Apr 29, 2019, at 7:13 AM, Chris Smith <chris at vspl.co.uk> wrote:
>>
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>>
>> Anybody out there know of any developments using CUDA cores for SDR? I imagine the effort required to do the functional decomposition to spread the processing over a number of cores would be immense. Would there be any advantage to such parallel processing as far as SDR was concerned? I haven’t seen any discussion on this subject on this reflector.
>>
>>
>