use Benchmark;
$s=($" x 1000)."test";
sub a { $_=$s;s/^ +//; }
sub b { $_=$s;while(index($_, " ") == 0) { $_ = substr($_, 1) }}

timethese(10000, {a=>\&a,b=>\&b});

Ergebnis:
Benchmark: 
timing 10000 iterations of
 a, b
...

         a:  0 wallclock secs ( 0.06 usr +  0.00 sys =  0.06 CPU) @ 166666.67/s (n=10000)

            (warning: too few iterations for a reliable count)

         b: 10 wallclock secs ( 9.18 usr +  0.00 sys =  9.18 CPU) @ 1089.32/s (n=10000)