Ok, I have done some tests with pperl (Persistent Perl), and the result was
a bit disappointing. :( Perhaps I am going about it the wrong way? Here is
what I did on my FreeBSD 4.7R system, running Perl 5.005_03.
-
#!/usr/bin/perl
use Benchmark;
timethese (1000, { 'old' => sub {
system ('/usr/bin/perl /temp/test.pl');
}, 'new' => sub {
system ('/usr/local/bin/pperl /temp/test.pl');
},
});
exit 0;
-
The "test.pl" script in my example is a huge Mail Delivery Agent program
that, though the full code is still present, has, for the test, an "exit 0"
at startup. The test-results were as follows:
-
Benchmark: timing 1000 iterations of new, old...
new: 82 wallclock secs ( 0.02 usr 0.09 sys + 8.55 cusr 15.03 csys =
0.00 CPU)
old: 125 wallclock secs ( 0.07 usr 0.07 sys + 110.34 cusr 13.97 csys
= 0.00 CPU)
-
As you can see, when it comes to compilation, pperl outperforms perl by a
factor of almost 2 (kinda). Not bad. But certainly not the reported tens to
hundreds times faster as has been reported. Should the difference not be
much greater?
Can someone tell me whether I am perhaps doing this the wrong way? It could
be that Perl is so smart, that is stops compilation when it encounters the
unconditional "exit 0" early on (which would account for a relatively small
difference in timings), but I doubt that.
Any suggestions will be appreciated.
- Mark
System Administrator Asarian-host.org
---
"If you were supposed to understand it,
we wouldn't call it code." - FedEx
___
Perl-Unix-Users mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
