You are viewing a plain text version of this content. The canonical link for it is here.
Posted to apache-bugdb@apache.org by Joseph Sloan <jj...@c-me.com> on 1997/03/28 00:00:02 UTC

general/278: 40-50% performance decrease relative to apache 1.1.3

>Number:         278
>Category:       general
>Synopsis:       40-50% performance decrease relative to apache 1.1.3
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    apache (Apache HTTP Project)
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Thu Mar 27 15:00:01 1997
>Originator:     jjs@c-me.com
>Organization:
apache
>Release:        1.2.x
>Environment:
Linux 2.0.29, Slackware 96 and RedHat 4.1
>Description:
To my dismay, every version of 1.2 I have compiled and installed turns
out to be 40-50% slower than 1.1.3 - I have been using the Zeusbench
program as a benchmark - whether I use 1 connection or 100 concurrent
connections, small files or large files, 1 request for the file or
several hundred requests, whether I request keep-alive or not, 1.2.x 
clocks in 40-50% slower than 1.1.3.
>How-To-Repeat:
Simply download zeusbench from www.zeus.co.uk/products/server
and compare apache 1.1 to 1.2 on any system 
>Fix:
Not yet - I'm amazed that nobody's mentioned this earlie
>Audit-Trail:
>Unformatted:



Re: general/278: 40-50% performance decrease relative to apache 1.1.3

Posted by Dean Gaudet <dg...@arctic.org>.
Sorry I've been working with Joseph on this one.  My suspicion is that his
benchmarks aren't very accurate -- too many concurrent requests, too few
children, too little time.  I'm still waiting for his response.

Dean

On Sun, 6 Apr 1997, Marc Slemko wrote:

> What happens if you add -DNO_LINGCLOSE to the EXTRA_CFLAGS line in your
> Configuration file, rerun Configure and recompile?  Does that help?
> 
> I do not see the same differences here, but I'm not running on Linux which
> may make a difference.
> 
> On Thu, 27 Mar 1997, Joseph Sloan wrote:
> 
> > 
> > >Number:         278
> > >Category:       general
> > >Synopsis:       40-50% performance decrease relative to apache 1.1.3
> > >Confidential:   no
> > >Severity:       serious
> > >Priority:       medium
> > >Responsible:    apache (Apache HTTP Project)
> > >State:          open
> > >Class:          sw-bug
> > >Submitter-Id:   apache
> > >Arrival-Date:   Thu Mar 27 15:00:01 1997
> > >Originator:     jjs@c-me.com
> > >Organization:
> > apache
> > >Release:        1.2.x
> > >Environment:
> > Linux 2.0.29, Slackware 96 and RedHat 4.1
> > >Description:
> > To my dismay, every version of 1.2 I have compiled and installed turns
> > out to be 40-50% slower than 1.1.3 - I have been using the Zeusbench
> > program as a benchmark - whether I use 1 connection or 100 concurrent
> > connections, small files or large files, 1 request for the file or
> > several hundred requests, whether I request keep-alive or not, 1.2.x 
> > clocks in 40-50% slower than 1.1.3.
> > >How-To-Repeat:
> > Simply download zeusbench from www.zeus.co.uk/products/server
> > and compare apache 1.1 to 1.2 on any system 
> > >Fix:
> > Not yet - I'm amazed that nobody's mentioned this earlie
> > >Audit-Trail:
> > >Unformatted:
> > 
> > 
> 
> 


Re: general/278: 40-50% performance decrease relative to apache 1.1.3

Posted by Marc Slemko <ma...@znep.com>.
What happens if you add -DNO_LINGCLOSE to the EXTRA_CFLAGS line in your
Configuration file, rerun Configure and recompile?  Does that help?

I do not see the same differences here, but I'm not running on Linux which
may make a difference.

On Thu, 27 Mar 1997, Joseph Sloan wrote:

> 
> >Number:         278
> >Category:       general
> >Synopsis:       40-50% performance decrease relative to apache 1.1.3
> >Confidential:   no
> >Severity:       serious
> >Priority:       medium
> >Responsible:    apache (Apache HTTP Project)
> >State:          open
> >Class:          sw-bug
> >Submitter-Id:   apache
> >Arrival-Date:   Thu Mar 27 15:00:01 1997
> >Originator:     jjs@c-me.com
> >Organization:
> apache
> >Release:        1.2.x
> >Environment:
> Linux 2.0.29, Slackware 96 and RedHat 4.1
> >Description:
> To my dismay, every version of 1.2 I have compiled and installed turns
> out to be 40-50% slower than 1.1.3 - I have been using the Zeusbench
> program as a benchmark - whether I use 1 connection or 100 concurrent
> connections, small files or large files, 1 request for the file or
> several hundred requests, whether I request keep-alive or not, 1.2.x 
> clocks in 40-50% slower than 1.1.3.
> >How-To-Repeat:
> Simply download zeusbench from www.zeus.co.uk/products/server
> and compare apache 1.1 to 1.2 on any system 
> >Fix:
> Not yet - I'm amazed that nobody's mentioned this earlie
> >Audit-Trail:
> >Unformatted:
> 
> 


Re: general/278: 40-50% performance decrease relative to apache 1.1.3

Posted by Dean Gaudet <dg...@arctic.org>.
Can you provide your Configuration and your configuration files, and the
Configuration you are using for 1.1.3?

Dean

On Thu, 27 Mar 1997, Joseph Sloan wrote:

> 
> >Number:         278
> >Category:       general
> >Synopsis:       40-50% performance decrease relative to apache 1.1.3
> >Confidential:   no
> >Severity:       serious
> >Priority:       medium
> >Responsible:    apache (Apache HTTP Project)
> >State:          open
> >Class:          sw-bug
> >Submitter-Id:   apache
> >Arrival-Date:   Thu Mar 27 15:00:01 1997
> >Originator:     jjs@c-me.com
> >Organization:
> apache
> >Release:        1.2.x
> >Environment:
> Linux 2.0.29, Slackware 96 and RedHat 4.1
> >Description:
> To my dismay, every version of 1.2 I have compiled and installed turns
> out to be 40-50% slower than 1.1.3 - I have been using the Zeusbench
> program as a benchmark - whether I use 1 connection or 100 concurrent
> connections, small files or large files, 1 request for the file or
> several hundred requests, whether I request keep-alive or not, 1.2.x 
> clocks in 40-50% slower than 1.1.3.
> >How-To-Repeat:
> Simply download zeusbench from www.zeus.co.uk/products/server
> and compare apache 1.1 to 1.2 on any system 
> >Fix:
> Not yet - I'm amazed that nobody's mentioned this earlie
> >Audit-Trail:
> >Unformatted:
> 
> 
>