andy pugh wrote: > On 31 May 2012 12:19, Michael Haberler <mai...@mah.priv.at> wrote: > > >> I think it is about time this community considers and decides what to do on >> RT network I/O devices - a plethora of special-case contraptions is >> undesirable IMO. >> > > I see a mention of a Realtime Ethernet interface in the page that > Chris Radek linked to. > That might be a good starting point for a common approach. > RTnet was designed for a bunch of nodes to communicate in a time-sliced fashion, requiring them to have custom ethernet interfaces that could send their responses in specific time slots. It has been said that this could be stripped down to make a simple two-node network that replies when ready.
I asked some people questions about latency and round-trip time on small packets, to see if it was feasible to use it to communicate with FPGA boards at a 1 ms update period. The scheme was for the PC to send a request, get position data compute new velocities and send them, so 3 messages per millisecond. People who had worked with rtnet thought this was unrealistic due to packet processing overhead. They thought it would take at least 300 us to create, send and extract a single packet. This seemed somewhat in line with ping times on general desktop systems, I was hoping that stripping down the interface would help. The other problem is that you need an RT-net driver for every ethernet chip/board you want to use. If the driver has not been written, then that chip/board cannot be used with RT-net. I think there are about 6-8 drivers in existence. Jon ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
