[USRP-users] X300 and UHD Converter

Michael West michael.west at ettus.com
Thu Apr 30 15:44:09 EDT 2015


Hi Serge,

At this time, the X300 only supports sc16 (be) over the wire.  All
conversion to and from that format is done on the host.  Doing other
formats on the wire are possible, but require UHD and FPGA changes.  It is
probably easiest to create optimized converters for the conversion to and
from sc16 (le).

Regards,
Michael

On Thu, Apr 30, 2015 at 12:17 PM, Serge Malo via USRP-users <
usrp-users at lists.ettus.com> wrote:

>  Hi all,
>
>  I have a performance issue with my PC connected to a X300.
> My Setup is: Windows 7 x64, X300 via 10GbE (intel X520-DA2), UHD 3.8.2
>
>  I was doing tests with the UHD benchmark_rate example, using both
> channels in Rx at 50Msps (100Msps in total)
> I did not get any overflows when using the default parameters (rx_cpu =
> fc32).
> However, when I use rx_cpu = sc16, I get overflows.
>
>  I looked into UHD, and found out that the converter from sc16_item32_be
> to fc32 is highly optimized (using SSE2), while sc16_item32_be to sc16
> (little-endian) is not optimized. I'm pretty confident that the overflows
> happen because of this (my CPU is not fast enough for this version of the
> converter)
>
>  My goal is to record 16-bit samples in little-endian format.
>  I have a few questions:
>  -Is there a way for the X300 to send samples over 10GbE in little-endian
> format, in order to avoid conversion on the Host CPU?
> -When using PCIe, can the X300 send samples in sc16_item32_le format otw?
> -In TX, when the application has 16-bit little-endian samples, does
> conversion happen on the Host CPU too?
>
>  Best regards,
>  Serge Malo
>
>
>
> _______________________________________________
> USRP-users mailing list
> USRP-users at lists.ettus.com
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20150430/84e3f271/attachment-0002.html>


More information about the USRP-users mailing list