None of those tests really work all that well with me. I have traffic shaping on my gateway machine.
Note: Depending on what port it is communicating on, it may be getting a lower priority (HTTP, and other traffic has higher priority), upload has a set limit, ACK packets get preference over all other packets.
Test was completed while bandwidth chart looked like (Graphs taken before bandwidth tests):
Looking at the graphs afterwards, it has to fight against other traffic (I live in a house with 7 other people) and as such we have to share the bandwidth, hence the bandwidth shaping. (HTTP has to stay fast, Torrents can slow down for essential traffic like IM's, email, http and other such things).
That being said, this is what our yearly transfer since we moved into the house looks like (Chart starts late, as I did not implement the graphing until about 5 weeks into it. Just imagine the line going down to 0):
Each of the dips is because the gateway was rebooted, or went offline when there was a power outage, or a stupid room mate who pulled the wrong plug
.