Overrun issue and packet loss

I am facing Overrun issue on 500K sampling rate with USRPN210. I see an
‘O’
after irregular time intervals when I use my flowgrpah . PC is directly
connected to USRP with Ethernet cable. PC details:

Processor: Intel Core i7-2600 CPU @ 3.40GHz 8
Memory: 3.8GiB

GNU C++ version 4.6.3; Boost_104601; UHD_003.005.000-26-gb65a3924
Gnruadio version: 3.6.3rc0

Also to check, I ran ./benchmark_rate --rx_rate 500e3 --duration 300,
three
times and following are the outputs:

Testing receive rate 0.500000 Msps
OO
Benchmark rate summary:
Num received samples: 149997045
Num dropped samples: 726
Num overflows detected: 2

Testing receive rate 0.500000 Msps

Benchmark rate summary:
Num received samples: 149998497
Num dropped samples: 0
Num overflows detected: 0

Testing receive rate 0.500000 Msps
OOO
Benchmark rate summary:
Num received samples: 149997408
Num dropped samples: 1089
Num overflows detected: 3

What could be the issue here? And what can I do to fix this issue as I
am
losing 3 to 4 packets in every 10k packets when I use my flow-graph. Any
comments would be appreciated.

Best Regards,

Sam

Sam, in linux you have to set something:

I think its this

net.core.rmem_max = 50000000

I tested the script on my box, there is no problem.

-Bob
PS:

[bob@localhost examples]$ ./benchmark_rate --rx_rate 500e3 --duration
30
linux; GNU C++ version 4.4.5 20101112 (Red Hat 4.4.5-2); Boost_104100;
UHD_003.002.001-release

Creating the usrp device with: …
– Opening a USRP2/N-Series device…
– Current recv frame size: 1472 bytes
– Current send frame size: 1472 bytes
Using Device: Single USRP:
Device: USRP2 / N-Series Device
Mboard 0: USRP-N210
RX Channel: 0
RX DSP: 0
RX Dboard: A
RX Subdev: WBX RX + Simple GDB
TX Channel: 0
TX DSP: 0
TX Dboard: A
TX Subdev: WBX TX + Simple GDB

Testing receive rate 0.500000 Msps

Benchmark rate summary:
Num received samples: 15001642
Num dropped samples: 0
Num overflows detected: 0
Num transmitted samples: 0
Num sequence errors: 0
Num underflows detected: 0


Sam mite wrote in post #1118824:

I am facing Overrun issue on 500K sampling rate with USRPN210. I see an
‘O’
after irregular time intervals when I use my flowgrpah . PC is directly
connected to USRP with Ethernet cable. PC details:

Processor: Intel Core i7-2600 CPU @ 3.40GHz 8
Memory: 3.8GiB

GNU C++ version 4.6.3; Boost_104601; UHD_003.005.000-26-gb65a3924
Gnruadio version: 3.6.3rc0

Also to check, I ran ./benchmark_rate --rx_rate 500e3 --duration 300,
three
times and following are the outputs:

Testing receive rate 0.500000 Msps
OO
Benchmark rate summary:
Num received samples: 149997045
Num dropped samples: 726
Num overflows detected: 2

Testing receive rate 0.500000 Msps

Benchmark rate summary:
Num received samples: 149998497
Num dropped samples: 0
Num overflows detected: 0

Testing receive rate 0.500000 Msps
OOO
Benchmark rate summary:
Num received samples: 149997408
Num dropped samples: 1089
Num overflows detected: 3

What could be the issue here? And what can I do to fix this issue as I
am
losing 3 to 4 packets in every 10k packets when I use my flow-graph. Any
comments would be appreciated.

Best Regards,

Sam

Try running it without anything else on the network or on your computer,
not even X11.

Andrew

can somebody please comment on it ? Thanks a head of time.


Sam

Andrew thanks for your reply.

I open terminal only (on ubuntu12.04 LTS), no other application/window.
Running the command again gave me following output.

Testing receive rate 0.500000 Msps
OOOOO
Benchmark rate summary:
Num received samples: 149995956
Num dropped samples: 1815
Num overflows detected: 5
Num transmitted samples: 0
Num sequence errors: 0
Num underflows detected: 0

Any suggestions? What is the reason for these overruns ?

On Mon, Aug 19, 2013 at 5:08 PM, Andrew D.
[email protected]wrote:

OO
Num overflows detected: 0
am losing 3 to 4 packets in every 10k packets when I use my flow-graph. Any


Discuss-gnuradio mailing list
[email protected]
Discuss-gnuradio Info Page

Best Regards,

Sam