After seeing that my wireless speeds were much faster than the speeds I was getting over Ethernet, I decided to invest in some new cables. I didn’t know it before, but I saw while I was changing them out that my current cables were Cat 5e. While putting my network together, I had just been grabbing whatever cables I could find in my scrap drawers. Now I have Cat 8 cables and my speeds jumped from 7MB/s to an average of over 40MB/s. It’s a much bigger improvement than I expected, especially for such a small investment.
Cat5e works fine for gigabit. If it’s not connecting at 1G, then the cable has been damaged and is probably connecting at 100M.
You should be seeing about 118MB/s in an iperf test on gigabit ethernet.
This. I’ve had issues at work while imaging classroom computers where some would finish in ~30 minutes and a few would need hours. All of the computers used Cat6 cables. This being a classroom, and students being absolute wankbags, they kept yanking the computers and kicking the cables, so the wires came loose from the plugs. I later used ethtool to debug the slow computers – the switch would only allow 10baseT link modes.
For later reference, the link light on most network cards is a different colour depending on link speed. Usually orange for 1G, green for 100M and off for 10M (with data light still blinking).
But that depends on the card. And some gigabit devices won’t do 10Mb at all.
True. Hence my caveat of “most cards”. If it’s got LEDs on the port, it’s quite likely to signal which speed it is at with those LEDs.
I haven’t yet come across a gigabit card that won’t do 10Mbit (edit: switches are a different matter) but sometimes I’ve come across cards that fail to negotiate speeds correctly, eg trying for gigabit when they only actually have a 4 wire connection that can support 100Mbit. Forcing the card to the “correct” speed makes them work.