On Saturday 06 June 2009 00:23:54 Alberto Trentadue wrote: > On Fri, 2009-06-05 at 12:14 -0700, Firas Abbas wrote: > > Hi, > > > > --- On Fri, 6/5/09, Patrick Strasser <patrick. > > stras...@tugraz.at> wrote: > > > Hello! > > > > > > ednet[1] and RaidSonic[2] sell boxes that can forward USB > > > ports over > > via the Linux USB/IP[3] system. > > > > Patrick > > > > From the web site the transfer rate of these boxes is > > 10/100Mb/s. So theoretically it should transfer a maximum of 12.5MByte/sec > which is far below the required 32Mbyte/sec of USRP1. > > Hello Firas. > Just for my understanding: > The 32Mbyte/s requirement is for a full speed - 4 channel USRP > operation. > But if we suppose a simpler application, e.g. 1 TX/RX channel only, > wouldn't this fit in the USB/IP?
This still on the decimation/interpolation rate. Each complex sample is 4 byte. Theoretically 100MBit Ethernet could sustain 3 MSample/s, so the decimation rate has to be 22 or higher (most probably substantially higher, as there is some overhead by the TCP(UDP?)/IP headers, but 32 (2MSample) should be fine). As Ethernet is full duplex, this is the number of samples/s that can be sustained per direction, in constrast to USB, where the bandwidth is shared. This will for sure add some latency, you will have to figure this out by yourself. Bottom Line: 1TX/1RX, 2 MSample/s, complex samples should be possible. Stefan -- Stefan Brüns / Bergstraße 21 / 52062 Aachen phone: +49 241 53809034 mobile: +49 151 50412019 _______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
