perf: improve authority state performance

[anonrig]

Thanks to @miguelteixeiraa and @lemire

Before

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Benchmark                            Time             CPU   Iterations        GHz cycle/byte cycles/url instructions/byte instructions/cycle instructions/ns instructions/url     ns/url      speed  time/byte   time/url      url/s
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BasicBench_AdaURL_With_Copy       2342 ns         2341 ns       298997    4.81719    13.6646    1003.73            64.271            4.70347         22.6575           4.721k    208.364 345.132M/s  2.89744ns   212.83ns 4.69858M/s
BasicBench_AdaURL_With_Move       2028 ns         2027 ns       346784      5.063    12.5322    920.545           58.1597            4.64083         23.4965         4.27209k    181.818 398.596M/s  2.50881ns  184.283ns 5.42643M/s

After

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Benchmark                            Time             CPU   Iterations        GHz cycle/byte cycles/url instructions/byte instructions/cycle instructions/ns instructions/url     ns/url      speed  time/byte   time/url      url/s
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BasicBench_AdaURL_With_Copy       2351 ns         2351 ns       300044    4.77282    13.7809    1012.27           64.1646            4.65604         22.2225         4.71318k    212.091 343.716M/s  2.90938ns  213.707ns 4.67931M/s
BasicBench_AdaURL_With_Move       2003 ns         2003 ns       349664     5.0045    12.3874    909.909           58.1361            4.69318          23.487         4.27036k    181.818 403.434M/s  2.47872ns  182.073ns  5.4923M/s

[lemire]

@anonrig This is now very good looking.

Can you have a look at what I did in PR #232 ? It is possible that you might be able to gain a little bit of extra performance by mimicking my crazy code. At the very least, you might enjoy it.

[anonrig]

The benchmark results after SWAR routine is like this.

Before

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Benchmark                            Time             CPU   Iterations        GHz cycle/byte cycles/url instructions/byte instructions/cycle instructions/ns instructions/url     ns/url      speed  time/byte   time/url      url/s
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BasicBench_AdaURL_With_Copy       2356 ns         2356 ns       295865    4.81269     13.896    1020.73           64.1745            4.61819         22.2259         4.71391k    212.091 342.922M/s  2.91611ns  214.202ns 4.66849M/s
BasicBench_AdaURL_With_Move       2053 ns         2050 ns       347598      5.026    12.4406    913.818           58.1733            4.67608          23.502         4.27309k    181.818 394.192M/s  2.53684ns  186.342ns 5.36647M/s

After

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Benchmark                            Time             CPU   Iterations        GHz cycle/byte cycles/url instructions/byte instructions/cycle instructions/ns instructions/url     ns/url      speed  time/byte   time/url      url/s
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BasicBench_AdaURL_With_Copy       2320 ns         2320 ns       301131    4.82366     13.677    1004.64           64.1993            4.69396         22.6421         4.71573k    208.273 348.337M/s  2.87078ns  210.872ns 4.74221M/s
BasicBench_AdaURL_With_Move       1998 ns         1998 ns       350259    5.09857    12.3552    907.545           58.1881            4.70961         24.0123         4.27418k        178 404.501M/s  2.47218ns  181.593ns 5.50682M/s