Minimizing or eliminating Overruns

Hello everyone,

Has anyone worked or is familiar with the file I am
currently running USRP1 to receive two streams each of 500,000 complex
samples/second from each of the two daughter boards (under the
dual_source_c) configuraton and dumping the stream straight to the
file_sink. Thus, I am reading a total of 8Mbytes/sec from the USB and
writing this to the file_sink. However, this is giving me a lot of
The overrun “uO” is printed in the file as below.

if (l_buffer->enqueue(l->buffer (), l_size) == -1)
fflush (stderr);

I believer this means that the USB is getting data from the USRP but is
dropping data as the l_buffer is not accepting it. The value l_size is
always 32768.

I am currently using a 2.26 GHz Intel Core 2 Duo Macbook with 2 GB (1067
MHz DDR3) RAM. I ran the file and the max.
throughput support I get is 32MBytes/sec. Therefore, streaming from the
USRP at 8MBytes/sec in itself is not a problem. However, after I connect
a file_sink, I get a lot of overruns.

Questions I would like to discuss:
a) Has anyone experienced writing at a rate more than 8MBytes/sec using
the file_sink. If so, what would be the best way to optimize this.
b) The file indicates that everytime there is a overrun I
lose 32768 bytes of data. However, I don’t know how many data samples
this would be after subtracting the overhead? Can anyone please help me
with this calculation.

Thank you very much,

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