I noticed that your figures are about an order of magnitude slower in absolute terms,

There appears to be nothing wrong with my set-up, and nothing I tried has allowed me to start and stop the regex engine in less than 1 microsecond.

I strongly suspect that all the timing showing millions/s are wrong. I think that the posters have tested code along the lines of:

use Benchmark qw[ cmpthese ];

my $str = '123456789';
cmpthese -1, {
 a=>q[ my ($a,$b) = $str =~  m/(23)[^8]+(8)/g; ], 
 b=>q[ my ($a,$b) = $str =~  m/(23)[^8]+(8)/; ],
 c=>q[ my ($a) = $str =~  m/(23)/g ],
 d=>q[ my ($a) = $str =~  m/(23)/; ],
};;
[download]

Note the my which means that when the tested are evaluated $str will be undefined.