Thanks Mark, The packetizer does look like it is what I need. Is there any documentation on it? I couldn't find any and all I get out of the packetizer is a high overflow for the duration of what is supposed to be the data out period. Or maybe I am not getting all my data in properly? Since I have a 1024 point FFT, each vector is 512 (even, odd). Then the result of all the cross-correlations, power, and power^2 calculations is a 768 bit by 512 channel data set that I want to fit into the 10 Gbe 64 bits at a time. A scope shows that I get nothing coming out of the packetizer's data port, and only a high overflow signal after every data input.
thanks, Sam On Thu, Aug 4, 2011 at 6:40 PM, Mark Wagner <[email protected]> wrote: > Hi Sam, > The packetizer block in the CASPER library buffers up data from a vector > accumulator and can send it out in bursts to the 10GbE block with the > appropriate data valid. It sounds like this may be what you need. > Mark > > On Thu, Aug 4, 2011 at 2:55 PM, Samuel Tun <[email protected]> wrote: >> >> Hello, >> >> I'm wondering if anyone has any examples or suggestions on how to get >> large numbers coming out of vector accumulators into the 10 Gbe block. >> This block takes 64 bits at a time, so I've tried to slice up whatever >> comes out fo the vacc into 64 bit slices, delaying subsequent ones so >> that mux selections lets me capture an entire spectral channel as it >> comes out of vacc. However, cycling through the mux means I miss all >> the channels in between count starts. Regardless, I may be going about >> it the wrong way, so any examples would be greatly appreciated. The >> ones I've seen concatenate up to 64 bits, so do not have this problem. >> If anyone wants to take a look, this is intended for getting out the >> data of a 3 antenna 500 MHz bandwidth correlator for the FASR Subsytem >> Testbed, FST. >> >> http://dl.dropbox.com/u/3531421/test1ghz10Gbe.mdl >> >> Thanks, >> Sam Tun >> > >
