Re: [SPOILER] Solution to Perl 'Expert' Quiz-of-the-Week #22

John J. Trammell wrote:
> On Mon, Aug 30, 2004 at 04:44:12AM -0400, Randy W. Sims wrote:
> [snip]
> 
>>Benchmark: timing 10 iterations of compound, simple...
>>compound: 866 wallclock secs (726.63 usr +  0.94 sys = 727.57 CPU) @ >  0.01/s (n=10)
>>simple:  4 wallclock secs ( 3.52 usr +  0.05 sys =  3.58 CPU) @ > 2.80/s (n=10)
>>         s/iter compound   simple
>>compound   72.8       --    -100%
>>simple    0.358   20246%       --
> 
> 
> That's odd--when I run that benchmark, I get about a factor of two:
> 
> Benchmark: timing 20 iterations of compound, simple...
> compound:  8 wallclock secs ( 7.70 usr +  0.06 sys =  7.76 CPU) @ 2.58/s (n=20)
> simple:  4 wallclock secs ( 4.13 usr +  0.05 sys =  4.18 CPU) @ 4.78/s (n=20)
>            Rate compound   simple
> compound 2.58/s       --     -46%
> simple   4.78/s      86%       --
> 
> That's on perl 5.6.1; on 5.8.4 the difference is a little more
> pronounced (more like a factor of 3).

Yeah, I just ran it on my Debian laptop w/ perl 5.8.4, and I get results 
similar to yours:

          s/iter compound   simple
compound   1.07       --     -55%
simple    0.484     120%       --

The original was run on Windows 2000 w/ perl 5.6.1. Similar processors, 
but run times are very different. I suspect the difference must be due 
to file caching. From observing the run on linux, it /seems/ like the 
entire word file is remaining in the file cache, where on Windows it 
reread the file each time. That's my best guess at what's going on, but 
it is just a guess.

Randy.