Randal L. Schwartz wrote:
>>>>>> "Peter" == Peter Farrar <[EMAIL PROTECTED]> writes:
>
>>> Replace
>>>
>>> EVIL: map { some;block;of;code;that;changes;$_ } @some_array;
>>>
>>> with
>>>
>>> GOOD: for (@some_array) { some;block;of;code;that;changes;$_ }
>
>
> The foreach loop *will* be faster.
>
for a small array, map is probably faster than foreach or for:
#!/usr/bin/perl -w
use strict;
use Benchmark;
timethese(999999,{_map=>sub{map{1} (1..10)},
_for=>sub{for(1..10){1}},
_foreach=>sub{foreach(1..10){1}}});
__END__
prints:
Benchmark: timing 999999 iterations of _for, _foreach, _map...
_for: 29 wallclock secs (22.80 usr + 0.02 sys = 22.82 CPU) @
43821.17/s (n=999999)
_foreach: 27 wallclock secs (23.01 usr + 0.04 sys = 23.05 CPU) @
43383.90/s (n=999999)
_map: 18 wallclock secs (18.39 usr + 0.01 sys = 18.40 CPU) @
54347.77/s (n=999999)
for a larger array(the following is done with array size of 100):
Benchmark: timing 99999 iterations of _for, _foreach, _map...
_for: 9 wallclock secs ( 8.15 usr + 0.00 sys = 8.15 CPU) @
12269.82/s (n=99999)
_foreach: 7 wallclock secs ( 8.16 usr + 0.00 sys = 8.16 CPU) @
12254.78/s (n=99999)
_map: 20 wallclock secs (14.77 usr + 0.00 sys = 14.77 CPU) @
6770.41/s (n=99999)
map is slow.
david
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
