even simpler bench_ip functions

[lemire]

This trims further the bench_ip code. The simpler code gives significant better performance in my tests.

Before:

parse_ip_std_fromchars                   :   0.81 GB/s   50.7 Mip/s  19.73 ns/ip   4.53 GHz  89.27 c/ip  490.53 i/ip   5.58 c/b  30.66 i/b   5.50 i/c 
parse_ip_fastfloat                       :   0.70 GB/s   44.1 Mip/s  22.70 ns/ip   4.52 GHz  102.65 c/ip  482.02 i/ip   6.42 c/b  30.13 i/b   4.70 i/c 

After:

parse_ip_std_fromchars                   :   1.14 GB/s   71.2 Mip/s  14.04 ns/ip   4.53 GHz  63.64 c/ip  381.33 i/ip   3.98 c/b  23.83 i/b   5.99 i/c 
parse_ip_fastfloat                       :   1.10 GB/s   68.8 Mip/s  14.53 ns/ip   4.53 GHz  65.81 c/ip  271.45 i/ip   4.11 c/b  16.97 i/b   4.12 i/c 

Both functions are about as fast with a small edge for the standard function.

[shikharish]

@lemire Can you also update CMakeLists.txt to fetch the latest counters commit?

[lemire]

@shikharish

Done. Merging.