why is iperf so much faster?

[dvc94ch]

// Thus the bandwidth with the maximum receive window would be:
    //
    // ((6291456*8) / (0,09/1000)) / 1000000 = 559_240 mBit/s
    //
    // An iperf run on localhost achieves around 60 gBit/sec:
    //
    // $ iperf -c 127.0.0.1
    // [  3]  0.0-10.0 sec  68.4 GBytes  58.8 Gbits/sec

and why is is so much faster than what you calculated would be the maximum possible? where does that calculation even come from?

[mxinden]

why is iperf so much faster?

Short answer: That is the question this repository should help answer in the future. Obvious things: Optimized for single socket, no encryption, no multiplexing, ...

why is is so much faster than what you calculated would be the maximum possible?

Likely because my calculation assumes that the receive window is the only buffer involved, while this is not the case in the real-world scenario of iperf.

where does that calculation even come from?

The calculation is based on the Bandwidth Delay Product.