RE: Re: each or keys?

Copying the hash means that keys and values might come out in different orders, and since you're quitting when the total reaches a certain point, you have a bogus benchmark.

I suggest setting all the values to be the same and rerunning the benchmarks.

-- Randal L. Schwartz, Perl hacker

Comment on RE: Re: each or keys?

Replies are listed 'Best First'.
RE: RE: Re: each or keys? by extremely (Priest) on Oct 11, 2000 at 04:00 UTC
Good call Benchmark: running each, keys, each for at least 10 CPU seconds... each: 11 wallclock secs (10.51 usr + 0.00 sys = 10.51 CPU) @ 417.32/s (n=4386) keys: 11 wallclock secs (10.53 usr + 0.00 sys = 10.53 CPU) @ 286.89/s (n=3021) Rate keys each keys 287/s -- -31% each 417/s 45% -- This is with `$top_hash{$i} = 50;` and `cmpthese(-10, { 'each' => \&e, 'keys' => \&k });` and otherwise the same as above. -- $you = new YOU; honk() if $you->love(perl)	[reply] [d/l] [select]

Replies are listed 'Best First'.

RE: RE: Re: each or keys?
by extremely (Priest) on Oct 11, 2000 at 04:00 UTC

Good call

Benchmark: running each, keys, each for at least 10 CPU seconds...
      each: 11 wallclock secs (10.51 usr +  0.00 sys = 10.51 CPU) @ 417.32/s (n=4386)
      keys: 11 wallclock secs (10.53 usr +  0.00 sys = 10.53 CPU) @ 286.89/s (n=3021)
      Rate keys each
keys 287/s   -- -31%
each 417/s  45%   --

This is with $top_hash{$i} = 50; and cmpthese(-10, { 'each' => \&e, 'keys' => \&k }); and otherwise the same as above.

--
$you = new YOU;
honk() if $you->love(perl)

[reply]
[d/l]
[select]