What happens to incoming data if a block is too slow?

Hi,
What happens to the incoming data from USRP, over the USB bus, when a
gnuradio block takes a lot of time to process its output? Is the data
buffered indefinitely or is it dropped as new samples come in case the
consumption rate is slow.

Thanks

GNURadio will indicates that overruns are occuring via the console. I
believe it is the u0 character.

Hi,
What happens to the incoming data from USRP, over the USB bus, when a
gnuradio block takes a lot of time to process its output? Is the data
buffered indefinitely or is it dropped as new samples come in case the
consumption rate is slow.

Thanks
That’s what’s called an “overrun”. Data gets dropped. The buffers are
large, but not infinite.


Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org

Is the data coming in the buffer FIFO? Is there a way to keep track of
how
much data is lost?

Is the data coming in the buffer FIFO? Is there a way to keep track of
how much data is lost?

Yes, the data are delivered FIFO.

In UHD, you can arrange for there to be timestamps on the data, so you
could look at the data and detect interruptions in the
monotonicity.

As long as overruns aren’t happening frequently (like more than once
every few dozen higher-layer-data-frames), then one can regard
such interrupts in data flow as equivalent to channel noise, which is
something you have to cope with regardless–even if your
sample stream is “perfect”.

In fact, if your channel model makes brash assumptions like the only
“noise” will be from the “natural physics” of the channel, then you’ve
already entered a state of sin. In general, radio channels aren’t
well-behaved with regard to “goop” on the channel, and I regard sampling
overruns, at a 10,000ft level, as broadly-equivalent to “channel
goop”.

Your mileage may vary. Void where prohibited. Batteries not included.
Some assembly required. No warantee as to fitness for a particular
purpose, etc, etc, etc :slight_smile:


Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org

On Mon, Jun 13, 2011 at 17:06, Marcus D. Leech [email protected]
wrote:

overruns, at a 10,000ft level, as broadly-equivalent to “channel goop”.

To a receiver, these appear as instantaneous phase jumps–which can, for
some modulations, take a long time to recover from. At least now with
UHD
one can update the receiver’s notion of sample number/timebase
accurately
when such a gap occurs.

Johnathan

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.

| Privacy Policy | Terms of Service | Remote Ruby Jobs