On Thu, Feb 07, 2002 at 09:57:48PM +0800, [EMAIL PROTECTED] wrote:
> From: "Sherwin Daganato" <[EMAIL PROTECTED]>
> > 'exponentiation' won. It's 52% faster than 'bit_shift' @ 500000
> > iterations.
> i dont think so :->
> [fooler@ns1 temp]$vi x.pl
> #!/usr/bin/perl
> $masklen = 24;
> for ( $i = 1000000; $i; $i--) {
> $mask = (2 ** ($masklen + 1)) - 1;
> }
> [fooler@ns1 temp]$time ./x.pl
> real 0m9.847s
> user 0m9.798s
> sys 0m0.018s
> [fooler@ns1 temp]$time ./x.pl
> real 0m9.831s
> user 0m9.821s
> sys 0m0.002s
> [fooler@ns1 temp]$vi x.pl
> #!/usr/bin/perl
> $masklen = 24;
> for ( $i = 1000000; $i; $i--) {
> $mask = (1 << ($masklen + 1)) - 1;
> }
> [fooler@ns1 temp]$time ./x.pl
> real 0m7.372s
> user 0m7.340s
> sys 0m0.010s
> [fooler@ns1 temp]$time ./x.pl
> real 0m7.351s
> user 0m7.346s
> sys 0m0.002s
I don't know. I can't explain it.
I just couldn't believe the way you benchmark Perl codes.
If that's the right way then why would anyone create the
Benchmark module (see CPAN :-).
This link might help explain things:
http://www.foo.be/docs/tpj/issues/vol3_3/tpj0303-0009.html
I'm raising the iteration to get >10s of cpu time.
root@emc
/tmp (#1093) cat testbench.pl
#!/usr/bin/perl -w
use strict;
use Benchmark;
my $masklen = 24;
my $r=Benchmark::cmpthese(10000000,
{
'bit_shift' => sub {
my $mask = (1 << ($masklen + 1)) - 1;
},
'exponentiation' => sub {
my $mask = (2 ** ($masklen + 1)) - 1;
}
});
exit;
root@emc
/tmp (#1094) w
11:07pm up 125 days, 14:48, 1 user, load average: 0.00, 0.10, 0.09
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
root pts/2 saturn.emc.com.p 10:12pm 0.00s 0.21s 0.01s w
root@emc
/tmp (#1095) ./testbench.pl
Benchmark: timing 10000000 iterations of bit_shift, exponentiation...
bit_shift: 15 wallclock secs (13.33 usr + 0.02 sys = 13.35 CPU) @
749063.67/s (n=10000000)
exponentiation: 12 wallclock secs (10.58 usr + -0.01 sys = 10.57 CPU) @
946073.79/s (n=10000000)
Rate bit_shift exponentiation
bit_shift 749064/s -- -21%
exponentiation 946074/s 26% -- # << STILL WON
root@emc
/tmp (#1098) cat bit_shift.pl exponentiation.pl
#!/usr/bin/perl
$masklen = 24;
for ( $i = 1000000; $i; $i--) {
$mask = (1 << ($masklen + 1)) - 1;
}
#!/usr/bin/perl
$masklen = 24;
for ( $i = 1000000; $i; $i--) {
$mask = (2 ** ($masklen + 1)) - 1;
}
root@emc
/tmp (#1099) w
11:19pm up 125 days, 15:00, 1 user, load average: 0.00, 0.03, 0.07
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
root pts/2 saturn.emc.com.p 10:12pm 0.00s 0.24s 0.02s w
root@emc
/tmp (#1100) time ./bit_shift.pl
1.67user 0.00system 0:01.67elapsed 99%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (260major+32minor)pagefaults 0swaps
root@emc
/tmp (#1107) w
11:22pm up 125 days, 15:03, 1 user, load average: 0.00, 0.02, 0.06
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
root pts/2 saturn.emc.com.p 10:12pm 0.00s 0.27s 0.01s w
root@emc
/tmp (#1108) time ./exponentiation.pl
1.61user 0.00system 0:01.60elapsed 100%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (260major+32minor)pagefaults 0swaps
--
$_=q:; # SHERWIN #
70;72;69;6e;74;20;
27;4a;75;73;74;20;
61;6e;6f;74;68;65;
72;20;50;65;72;6c;
20;6e;6f;76;69;63;
65;27;:;;s=~?(..);
?=pack q$C$,hex$1;
;;;=ge;;;;;eval;;;
_
Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph
To leave: send "unsubscribe" in the body to [EMAIL PROTECTED]
To subscribe to the Linux Newbies' List: send "subscribe" in the body to
[EMAIL PROTECTED]
