On Fri, Oct 07, 2005 at 10:24:28AM -0400, Rutherdale, Will wrote:
> Thanks, Tim and Jonathan.
> 
> The motivator for the question is that I'm looking to recommend upgrades to
> management in a month or two, e.g. for library versions (DBI, DBD, etc.),
> Perl versions, and extra libraries to install.  These upgrades would be
> applied to a number of machines so they have to be planned.
> 
> I'm not sure what you meant, Tim, when you referred to extensions.  Do you
> mean non-pure-Perl libraries that require compiling C code, such as DBI and
> most DBDs?

Yes.

> If so I can see the sense in the restriction, as usually the
> library has to be compiled so as to be compatible with the Perl that will be
> using it.
> 
> I'm wondering if anyone has benchmarks comparing Perl with threads against
> Perl without threads.  If it's a 2% slowdown (for instance) it wouldn't be
> such a big deal, but 50% would bother me.

At least 10%, I've often seen 30% mentioned, and sometimes 100%+.

As always, you should do your own benchmarks for your own tasks.

Tim.

> -Will
> 
> BTW  My personal take on email layout is that top-quoting is better 90% of
> the time because people see up front what I have to say uncluttered and can
> refer to the earlier discussion as needed.  I use bottom quoting when I plan
> to respond point-by-point several times.
> 
> -----Original Message-----
> From: Tim Bunce [mailto:[EMAIL PROTECTED] 
> Sent: Friday 07 October 2005 04:39
> To: Jonathan Leffler
> Cc: Rutherdale, Will; List - DBI users
> Subject: Re: Perl DBI and threads.pm
> 
> 
> On Thu, Oct 06, 2005 at 07:42:14PM -0700, Jonathan Leffler wrote:
> > On 10/6/05, Rutherdale, Will <[EMAIL PROTECTED]> wrote:
> > 
> > > We have some Perl DBI scripts and have been running them on a Perl
> > > compiled
> > > with threading disabled. (I.e. perl -V gives ...usethreads=undef...). I
> > > also have another environment where Perl is compiled with
> > > usethreads=define
> > > and have found the threading capabilities useful. (This is the Perl 5.8
> > > threads.pm <http://threads.pm> variety.)
> > >
> > > My question is this: if I were to upgrade the platform with the
> threadless
> > > Perl to a threaded Perl, would this risk reducing stability or
> reliability
> > > of the existing Perl DBI applications?
> > >
> > > These applications wouldn't be using multithreading, just switching to a
> > > Perl with threading enabled.
> > 
> > I don't have the hands on experience, so if someone with such experience
> > contradicts me, you should probably take their advice over mine.
> > 
> > My understanding is that a Perl with multi-threading where you do not use
> > the multithreading should be as stable as a Perl without the
> multithreading.
> > However, there's a chance that some modules may not be so happy. OTOH, DBI
> > shouldn't be one of those modules; it ensures single-tracking through it.
> > OTOOH, if you aren't using the multi-threading, there still shouldn't be
> any
> > problem.
> 
> I'd only add that a) the two builds aren't binary compatible so you
> can't just copy extensions between them, and b) perl built with
> multi-threading enabled is significantly slower.
> 
> Tim.
> 
> 
>      - - - - - - -  Appended by Scientific-Atlanta, Inc.  - - - - - - -  
> This e-mail and any attachments may contain information which is 
> confidential, proprietary, privileged or otherwise protected by law. The 
> information is solely intended for the named addressee (or a person 
> responsible for delivering it to the addressee). If you are not the intended 
> recipient of this message, you are not authorized to read, print, retain, 
> copy or disseminate this message or any part of it. If you have received this 
> e-mail in error, please notify the sender immediately by return e-mail and 
> delete it from your computer.
> 

Reply via email to