> I'll try to get you more info, but can you try to reproduce it there? I've been running this testing for several hours now. Each time, once I stop cycling the ports down and up, the performance recovers nicely. -- MST