[USRP-users] UHD | X300 - ERROR_CODE_TIMEOUT with low rates

Guilbert, Kyle J kguilbert at mitre.org
Wed Nov 15 12:19:15 EST 2017


I am experiencing a similar timeout problem to the one reported in the thread "UHD 3.10.2 | X300 - High CPU load even for low samples rate". 
I'm using an X310 with LFRX/LFTX and UHD (maint) and can reproduce the issue with benchmark_rate for RX and TX somewhere under 10 MS/s.
When I use benchmark_rate with RX only as suggested by Michael West, there's no issue. But when I do simultaneous RX and TX I get a flurry of ERROR_CODE_TIMEOUT messages after a few seconds:

benchmark_rate --rx_rate 20e6 --tx_rate 20e6 --duration 20: OK
benchmark_rate --rx_rate 10e6 --tx_rate 10e6 --duration 20: OK
benchmark_rate --rx_rate 5e6 --tx_rate 5e6 --duration 20: ERROR_CODE_TIMEOUT errors
(failures continue for lower rates)

benchmark_rate --rx_rate 20e6 --duration 20: OK
benchmark_rate --rx_rate 10e6 --duration 20: OK
benchmark_rate --rx_rate 5e6 --duration 20: OK
benchmark_rate --rx_rate 1e6 --duration 20: OK
(note: the same goes for tx only)

Also I can confirm that when I change the timeout value in the get_recv_buff() call in device3_io_impl.cpp from 1 us to 10 ms, 
all these configurations work. Let me know if there's any more information I can provide.


More information about the USRP-users mailing list