From: Roger Pack <rogerpack2005@...>
Date: 2008-09-27T05:30:36+09:00
Subject: Re: Hash optimization question

> Aside: ruby-1.9 keeps links between hash entries, exactly for the 
> purpose of faster iteration. It also has the potentially useful 
> side-effect that hashes iterate in the same order that the entries were 
> inserted.

Interesting.
It appears that whatever the internal linking mechanism, it still 
differs from Array#each speed-wise:

1.8.6 Hash:

>> Benchmark.measure { a = {:a => :b, :c => :d}; 100_000.times { a.each {} }}
=> #<Benchmark::Tms:0x26c2f8 @total=0.17, @cutime=0.0, @label="", 
@stime=0.0, @real=0.173249959945679, @utime=0.17, @cstime=0.0>

1.8.6 Array:
>> Benchmark.measure { a = [[:a => :b], [:c => :d]]; 100_000.times { a.each {} }}
=> #<Benchmark::Tms:0x5de7c4 @total=0.07, @cutime=0.0, @label="", 
@stime=0.0, @real=0.0752890110015869, @utime=0.07, @cstime=0.0>

so 0.17 to 0.7 s

and on 1.9 Hash:
>> Benchmark.measure { a = {:a => :b, :c => :d}; 100_000.times { a.each {} }}
=> #<Benchmark::Tms:0x130efc @label="", @real=0.142009973526001, 
@cstime=0.0, @cutime=0.0, @stime=0.0, @utime=0.14, @total=0.14>

1.9 Array:
>> Benchmark.measure { a = [[:a, :b], [:c, :d]]; 100_000.times { a.each {} }}
=> #<Benchmark::Tms:0x13379c @label="", @real=0.0471560955047607, 
@cstime=0.0, @cutime=0.0, @stime=0.0, @utime=0.05, @total=0.05>

so 0.14 to 0.05 s

The numbers seem very similar, comparison wise, to each other.

-=R
-- 
Posted via http://www.ruby-forum.com/.