We routinely use 3m cables. We have a few 5 m cables in use, but not for GUPPI production use, just in the lab. But they seem to work fine. I remember that the ibob would not drive a very long cable...
John > > Hi Billy, > > By in the lab I was trying to say I have a box of 3m cables sitting unused > on the floor. But I just checked... and the PAPER project is running > with 8 Roach1s each using 1 3m passive cable to send data to/from a > Fujistu 10GbE switch. I think the user data rates are 6.4 Gbps each way. > > Matt > > > On Sun, 4 Mar 2012, Barott, William Chauncey wrote: > >> Thanks, Dan and Matt- hearing that 3m is being used is enough for me. >> FWIW I'll be running these to CX4 modules on a Fujitsu XG0448 for most >> of the time. I'll take a shot at 3m. >> >> Billy >> >> -----Original Message----- >> From: [email protected] >> [mailto:[email protected]] On Behalf Of Matt Dexter >> Sent: Sunday, March 04, 2012 1:35 PM >> To: casper >> Subject: Re: [casper] Any cable length considerations for ROACH? >> >> Hi, >> >> most of my knowledge, such as it was, is probably obsolete by now. >> >> I haven't performed the tests to be able to say, one way or another, >> how the Roach1 nor Roach2 w/ long passive cable assemblies. >> Sorry - just haven't had the time. >> >> The longest passive cable I have in the lab is 3m. So I could set >> up at least a crude go/nogo test with those. For longer cables >> someone would have to provide the cable. >> >> But what are the two points of particular interest ? >> Roach1-a to/from Roach1-b ? >> Roach1 to/from commercial 10GbE switch (Fujitsu most likely) >> Roach1 to/from commercial 10GbE NIC from Myricom or Checksio or ... >> Roach1 to (only to) iBob or BEE2 or some legacy system ? >> same sort of idea for Roach2 w/ CX4 mezzanine card. >> >> Matt Dexter >> >> >> On Sun, 4 Mar 2012, Dan Werthimer wrote: >> >>> >>> hi billy, >>> info on copper and optical cables we've tested is at: >>> https://casper.berkeley.edu/wiki/Equipment_Cables >>> >>> virtex2 fpga's (ibob, bee2) do not meet 10GbeCX4 specifications, >>> but virtex5 fpga's (roach) does meet the 10GbeX4 spec, >>> so 5 meter CX4 cables should work. >>> >>> the longest cable we've used is 3 meters. >>> >>> if you want to go longer than 5 meters, there are several fiber optic >>> cables >>> you can use, with optical transceivers built into the CX4 connectors. >>> >>> matt dexter is the expert here. >>> >>> best wishes, >>> >>> dan >>> >>> On Sun, Mar 4, 2012 at 9:07 AM, Barott, William Chauncey >>> <[email protected]> wrote: >>> >>> Hi all, >>> >>> >>> >>> Are there any cable length considerations for the ROACH? The >>> BEE2s had >>> some limitations of not playing well > 2m copper CX4 cables >>> (discussed in >>> past threads). I can't find any threads on similar issues with >>> the ROACH - >>> does that mean there are none? What lengths are folks >>> successfully using >>> for copper XAUI and/or 10GbE? >>> >>> >>> >>> Billy >>> >>> >>> >>> >>
