Is it really more efficient?

I got the impression that 'efficient' was more to do with line consumption than CPU cycles, and even if it is do with CPU cycles, map should be just as fast as a foreach if not faster in this situation. Let's see ...

use strict;

use Benchmark 'cmpthese';

my @y_vals = my @x_vals = 1 .. 50;
cmpthese(-10, {
  map => sub {
    return map "($x_vals[$_],$y_vals[$_])", 0 .. $#x_vals
  },
  for => sub {
    my @ret;
    $ret[$_] = "($x_vals[$_],$y_vals[$_])"
      for 0 .. $#x_vals;
    return @ret;
  },
});

__output__

Benchmark: running for, map, each for at least 10 CPU seconds...
       for: 11 wallclock secs (10.19 usr +  0.00 sys = 10.19 CPU) @ 75
+36.80/s (n=76800)
       map: 11 wallclock secs (10.34 usr +  0.01 sys = 10.35 CPU) @ 77
+68.31/s (n=80402)
      Rate for map
for 7537/s  -- -3%
map 7768/s  3%  --
[download]

Ever so slightly faster, but essentially negligble. As always, take with a grain of salt. But I'd say that's about as quick as your going to get without going down the dark path of optimization.
HTH

_________
broquaint