Re: Perl Quiz of the Week #23

On Fri, 03 Sep 2004 08:24:53 -0400, colin.rafferty-/PgpppG8B+R7qynMiXIxWgC/[email protected]
<colin.rafferty-/PgpppG8B+R7qynMiXIxWgC/[email protected]> wrote:
> Xavier Noria wrote:
> > On Sep 3, 2004, at 14:00, colin.rafferty-/PgpppG8B+R7qynMiXIxWgC/[email protected] wrote:
> 
> >> paias746 $ time ~/misc/qotw/parens 12 | head
> >> (((((((((((())))))))))))
> >> ((((((((((()()))))))))))
> >> (((((((((()(()))))))))))
> >> ((((((((()((()))))))))))
> >> (((((((()(((()))))))))))
> >> ((((((()((((()))))))))))
> >> (((((()(((((()))))))))))
> >> ((((()((((((()))))))))))
> >> (((()(((((((()))))))))))
> >> ((()((((((((()))))))))))
> >>
> >> real 0m0.05s
> >> user 0m0.01s
> >> sys  0m0.01s
> 
> > My goodness:
> 
> > % time perl balanced_gen.pl 12
> > perl balanced_gen.pl 12  4.38s user 0.03s system 83% cpu 5.255 total
> 
> You are looking at the wrong number.  In the one you're showing, I'm
> doing `head', so it's cut off after 10 results.  You really should be
> looking at the results of the full run to see that we're in the same
> ballpark.

    I assert that most of the time for any non-brain-damaged
implementation will be spent in the tty driver.  In my implementation
(which just so happens to produce its output in lexicographic order):

    % time parens 12
    [ output not shown ]
    real    0m18.137s
    user    0m5.940s
    sys     0m3.912s
    % time parens 12 | head
    [ output not shown ]
    real    0m0.079s
    user    0m0.050s
    sys     0m0.004s
    % time parens 12 | tail
    [ output not shown ]
    real    0m4.780s
    user    0m4.510s
    sys     0m0.114s
    % time parens 12 > /dev/null
    real    0m4.541s
    user    0m4.460s
    sys     0m0.012s

    The "tail" version has to generate all the output, it just doesn't
have to display most of it to the s*l*o*w tty.  Interestingly, the
"/dev/null" version isn't that much faster, which reinfoces the idea
that the slowness isn't in the I/O subsystem per se, but rather in the
tty driver specifically.

    - Kurt