comparison between our to_chars and dragonbox's to_chars

[jaja360]

Following the discussion in #24, I added a comparison of our to_chars and dragonbox's to_chars implementation.

Here are some results on my computer (Ryzen 9 9900x) :

❯ ./build/benchmarks/benchmark -a "just_string" -f ./data/canada.txt
number type: binary64 (double)
# read 111126 lines
just_string_ours              :  2813.98 MB/s (+/- 0.7 %)     2.09 MB      6.68 ns/f   149.68 Mfloat/s
                                    9.90 i/B   186.20 i/f (+/- 0.0 %)      1.97 c/B    37.07 c/f (+/- 0.2 %)
                                    5.02 i/c      0.00 b/f                 0.00 bm/f      5.55 GHz
just_string_dragonbox         :  4338.15 MB/s (+/- 1.2 %)     2.09 MB      4.33 ns/f   230.75 Mfloat/s
                                    5.65 i/B   106.28 i/f (+/- 0.0 %)      1.28 c/B    24.01 c/f (+/- 0.2 %)
                                    4.43 i/c      0.00 b/f                 0.00 bm/f      5.54 GHz

[lemire]

Wow. This is a massive find !!! And unexpected. Great work!!!