Ethernet

I was wondering if those who implemented 100BaseT ports in V2 Pro would want to share their experiences? I would like to know how difficult it was overall, what unexpected obstacles you had to go through and what kind of performance in terms of sustained TCP packet rate has been achieved?

Thanks, /Mikhail

Reply to
MM
Loading thread data ...

I once worked on a design that had sixteen 100BaseT ports in a pair of Virtex-E 2000. That used an external multiport MAC and PHY though, but did manage to achieve 100% of the theoretical maximum performance for IP.

TCP performance depends a lot on latency. You should try to keep latency low. I assume that some of your protocol stack will be in software. Offload as many functions into hardware as you can, e.g. IP header checksum calculations, etc.

Also, process buffers in place to avoid the overhead of copying them (or if you want to copy buffers, use hardware acceleration to do the work, e.g. DMA).

Regards, Allan

Reply to
Allan Herriman

Only if your buffers/windows are too small.

Standard party line is that you need a window/buffer big enough to cover the round trip time. That includes both speed of light delays and processing times at both ends.

formatting link
formatting link

You need the window scale option to get interesting buffer sizes. RFC 1323

That's not to say that you shouldn't minimize latency.

--
The suespammers.org mail server is located in California.  So are all my
other mailboxes.  Please do not send unsolicited bulk e-mail or unsolicited
 Click to see the full signature
Reply to
Hal Murray

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.