Real World Throughput Test vs Published#s vs Estimation

Has anyone validated throughput tests with load on:
hex
RB450Gx4
RB4011
CCR1009

Please add your actual results to the mix.

My estimation process is as follows to mimic real world based on my limited experience.
A. 512 byte
B. Mpbs
C. Range Weighting 1 (lower) 25simple queue rules + 25 filter rules weighted 1/3 and 2/3 respectively
D. Range Weighting 2 (upper) 25simple queue rules + 25 filter rules weighted 2/3 and 1/3 respectively

For example
Hex: Range 502 - 618
RB4011: Range 3135-3710
CCR1109: Range 4266-5720

Too conservative? Is upper number actually accurate enough by itself?

With a little bit of luck, soon I will be able to test an hEX, with gigabit and PPPoE. Answering here to follow the topic.

Thanks paternot, looking forward.
I have no way of testing the other units at such high throughputs, would take someone with 3-5 gig wan lines or something

Just to say I didn’t forget this. It’s just I was offered to be one of the beta testers of the new gigabit plan by my ISP - but things are stopped now. Weird, I really don’t know why.