Report standard deviation of the speedup ratio

[Shnatsel]

Report standard deviation of the speedup ratio.

Here are results for running the same Rust binary that exits immediately twice:

Before:

  'target/release/examples/decode' ran
    1.03x faster than 'target/release/examples/decode'

After:

  'target/release/examples/decode' ran
    1.03x ± 0.65 faster than 'target/release/examples/decode'

Results from a real-world benchmarking run:

  'flac -fd testsamples/extra/*.flac' ran
    1.56x ± 0.08 faster than 'target/release/examples/decode testsamples/extra/*.flac'

I'd also really like to see statistical significance estimation and getting rid of the word "faster" and all that green in case it's not, but I don't have the time to work on it right now.

[sharkdp]

Thank you very much - this looks great!

My only concern is readability. The ran 1.03x ± 0.65 faster than doesn't read like an English sentence anymore if you pronounce x as "times". Maybe:

    1.03 ± 0.65 times faster than

?

[Shnatsel]

I've updated the code with the message you suggested.

[sharkdp]

Thank you!