use Benchmark qw( cmpthese );

cmpthese(
-3, {

del => sub {
    my @B = (1..100);
    delete @B[0,-1];
},

shiftNpop => sub {
    my @B = (1..100);
    shift @B;
    pop @B;
},
sliceAssi => sub {
    my @B = (1..100);
    @B = @B[1..$#B-1];
},

splice => sub {
    my @B = (1..100);
    splice @B,0,1;
    splice @B,-1;
}
});
__END__
Benchmark: running del, shiftNpop, sliceAssi, splice, each for at least 3 CPU seconds...
       del:  4 wallclock secs ( 3.08 usr +  0.00 sys =  3.08 CPU) @ 67710.85/s (n=208414)
 shiftNpop:  3 wallclock secs ( 3.06 usr +  0.00 sys =  3.06 CPU) @ 68042.44/s (n=208414)
 sliceAssi:  3 wallclock secs ( 3.20 usr +  0.00 sys =  3.20 CPU) @ 18594.44/s (n=59558)
    splice:  3 wallclock secs ( 3.09 usr +  0.00 sys =  3.09 CPU) @ 67360.70/s (n=208414)
             Rate sliceAssi    splice       del shiftNpop
sliceAssi 18594/s        --      -72%      -73%      -73%
splice    67361/s      262%        --       -1%       -1%
del       67711/s      264%        1%        --       -0%
shiftNpop 68042/s      266%        1%        0%        --