NT/IIS/PerlEx vs ASP : stupid benckmark

[Valter Mazzola <tx...@hotmail.com>]

>From: Gunther Birznieks <gu...@extropia.com>
>To: soulhuntre@soulhuntre.com
>CC: modperl@apache.org
>Subject: Re: mod_perl compat on win32? RE: mod_perl weaknesses? help me 
>build a case....
>Date: Thu, 06 Apr 2000 13:40:51 +0800
>
>
>
>Soulhuntre wrote:
>
> > Hiya :)
> >
> > > -----Original Message-----
> > > From: Leslie Mikesell [mailto:les@mcs.net]
> > > Sent: Wednesday, April 05, 2000 1:34 PM
> > > To: soulhuntre@soulhuntre.com
> > > Cc: modperl@apache.org
> > > Subject: Re: mod_perl weaknesses? help me build a case....
> > >
> > > Your problem here is going to be that mod_perl is not thread-safe
> > > and will serialize everything when running under the threaded
> > > model that apache uses under windows. If your scripts are fast enough
> > > you might be able to live with this if you use it as a back end
> > > to a lightweight front-end proxy which a busy site needs anyway.
> >
> > Ok... that makes sense :)
> >
> > On a side note... is mod_perl/apache on linux compatible with 
>IIS/operlex
> > from activestate on the Win32 side?
> >
>
>For all intents and purposes, PerlEx is basically compatible with
>mod_perl/Apache::Registry. However, the END {} blocks are interpreted
>differently (augh!)... which shouldn't really be that much of an issue for
>most programs.
>
>The good thing about PerlEx as opposed to mod_perl/Win32 is that instead of
>serializing all the calls to one Perl interpreter, ActiveState Perl 
>actually
>creates multiple Perl interpreter "objects" that can be round-robined among
>the IIS threads.
>
> >
> > If so, that's a big boost for me...
> >
>
>I've been testing all my code on both PerlEx and mod_perl. And haven't had
>very many problems at all with compatibility. I think UNIX/mod_perl is 
>faster
>than NT/IIS/PerlEx on the same hardware (from my unscientific observation),
>but PerlEx seems much faster than ASPs.
>

i've made a stupid unscientific benckmark:

the program loops 1000000 and print a series of "a ", PerlEx takes the same 
time as ASP (same NT machine) , BUT processor goes 100% with PerlEx, 45% 
with ASP.

Can someone benchmark mod_perl under Win32, using the same stupid program ?



>Later,
>    Gunther
>
>

______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com

Re: NT/IIS/PerlEx vs ASP : stupid benckmark

[Gunther Birznieks <gu...@extropia.com>]

It may be a "stupid" benchmark. But no one seems to have commented on the CPU
rates. Why was PerlEx 100% and PerlScript 45% on the same machine, same
ActiveState Perl (presumable), same CPU config. And yet took the same amount of
time to complete.

I find that interesting. I suspect that it is a trick with how the OS views CPU
time (eg user time vs system calls vs IO wait) in the two architectures, but it
would be interesting to know why this is. Especially if mod_perl ends up
adopting a similar round robining of Perl interpreters among apache threads
later on down the line (becoming more similar to PerlEx architecture).

Later,
   Gunther

Nicolas MONNET wrote:

> On Thu, 6 Apr 2000, Valter Mazzola wrote:
> |i've made a stupid unscientific benckmark:
> |
> |the program loops 1000000 and print a series of "a ", PerlEx takes the same
> |time as ASP (same NT machine) , BUT processor goes 100% with PerlEx, 45%
> |with ASP.
> |
> |Can someone benchmark mod_perl under Win32, using the same stupid program ?
>
> I don't mean to be rude, but this is one stupid benchmark! Basically
> useless for that matter. You're not going to demonstrate anything with
> this.
>
> Now a good question is: what would be a good benchmark?
>
> What about doing some real life stuff, like get big results from a
> database, and calculate something over them, and print the (big?) result
> back?
>
> (Now this is not flamebait, I'm really wondering: why run mod_perl apps on
> WinNT? )

Re: NT/IIS/PerlEx vs ASP : stupid benckmark

[Nicolas MONNET <ni...@monnet.to>]

On Thu, 6 Apr 2000, Valter Mazzola wrote:
|i've made a stupid unscientific benckmark:
|
|the program loops 1000000 and print a series of "a ", PerlEx takes the same 
|time as ASP (same NT machine) , BUT processor goes 100% with PerlEx, 45% 
|with ASP.
|
|Can someone benchmark mod_perl under Win32, using the same stupid program ?

I don't mean to be rude, but this is one stupid benchmark! Basically
useless for that matter. You're not going to demonstrate anything with
this. 

Now a good question is: what would be a good benchmark?

What about doing some real life stuff, like get big results from a
database, and calculate something over them, and print the (big?) result
back?

(Now this is not flamebait, I'm really wondering: why run mod_perl apps on
WinNT? )