[hpsdr] openHPSDR at the forefront of SDR development

Hermann hvh.net at gmail.com
Tue Aug 26 10:40:02 PDT 2014


Dear John, all,

thank you all for your thoughts. Here's my 2-cent:

the discussion focus too much on Hardware, and the Jetson board is surely
an experimental platform. It doesn't makes much sense to ponder upon if
this board has one or two ethernet interfaces. Of course, this does matter
if a certain architecture should be tested. Btw, the upcoming Tablets and
Handhelds will have chips like the Tegra K1, and in a couple of years even
more powerful devices. So, we will indeed have handhelds with
super-computing power - at least in a certain sense.

Hardware is subject to fast changes, but how do we keep up with the
Software? Why do you think is Altera, e.g., providing an interface to
OpenCL? Nvidia is doing a really great job in providing not only fast
hardware, but also a complete programming model, together with all
necessary tools. In a completely different domain (but also embedded), chip
providers are building great hardware, with 6 or more cores etc etc., but
don't tell the software architects how to program these fine controllers,
or how to migrate software, which grew out of 20 years of development (and
as such is invaluable). The 'multicore problem' is still not solved (some
may claim they have), and the majority of algorithms and code (with some
very few exceptions, like FFTs, or matrix multiplication) is NOT easily
parallelized.

Why does nobody (with very few but remarkable exceptions) take existing
code (KISS Console, PowerSDR, cuSDR) and work on it to implement this or
that feature? I receive so many emails and comments on the mailing lists,
asking for new features. And if it's not going fast enough, people start
declaring cuSDR as vaporware, hi. It is not much more difficult, than, as
Joe, K5SO has reported so wonderful, start working on code for FPGAs.

Hardware changes fast, Software not. The idea of open source is not only
that the software is free. Much more important is that a lot of people
start contributing to it.

For me this is the main message from Phil's original post.

73, Hermann
DL3HVH
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20140826/590a6ef0/attachment-0003.htm>


More information about the Hpsdr mailing list