It really shouldn't surprise anyone that Open Source software has issues on a Windows box, especially 2000 which is running on a 6 year old design. The question might be would the same issue be there if you were running MSSQL and .NET instead of PHP and MySQL? Also, 512 is a lot of RAM for LAMP but its only a light snack for a Windows server. That's our minimum standard for office workstations. Still, for a fairer comparison of technology (not hardware) I would double the RAM, install 2003, and eliminate all non-essential services. I also might try it turning of page files and going straight RAM.
Michael Rhoadarmer, Media Systems Manager Wheaton College Wheaton, IL 60187 michael.r.rhoadarmer@wheaton.edu
tthompson+mediawiki@envisionware.com 06/13/2005 12:20:11 AM >>>
-----Original Message----- From: mediawiki-l-bounces@Wikimedia.org [mailto:mediawiki-l-bounces@Wikimedia.org]On Behalf Of FL Sent: Sunday, June 12, 2005 3:02 PM To: MediaWiki announcements and site admin list Subject: Re: [Mediawiki-l] Windows Apache & IIS vs Linux Apache performancedifferences
Question: is this comparison between LAMP vs WIMP implementations of the same software being made on molecularly identical hardware? The LAMP configuration is about 3 1/2 times faster than the WIMP configuration.
The comparison is on vastly different hardware. The LAMP hardware is a Pentium 3 laptop at 850Mhz, 200MB RAM. The WIMP hardware is a Pentium 4 server at 2+ Ghz, 512MB RAM.
This actually makes the speed of the LAMP implementation even more stunning; at half the CPU and half the RAM, it's consistently 4 times faster than the WIMP server.
- Troy Thompson
_______________________________________________ MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
On 6/13/05, Michael Rhoadarmer Michael.R.Rhoadarmer@wheaton.edu wrote:
It really shouldn't surprise anyone that Open Source software has issues on a Windows box, especially 2000 which is running on a 6 year old design. The question might be would the same issue be there if you were running MSSQL and .NET instead of PHP and MySQL? Also, 512 is a lot of RAM for LAMP but its only a light snack for a Windows server. That's our minimum standard for office workstations. Still, for a fairer comparison of technology (not hardware) I would double the RAM, install 2003, and eliminate all non-essential services. I also might try it turning of page files and going straight RAM.
Michael Rhoadarmer, Media Systems Manager Wheaton College Wheaton, IL 60187 michael.r.rhoadarmer@wheaton.edu
Why is building an even bigger server than the one already twice as fast a
"fairer" comparison? How much does the LAMP configuration have to be handicapped before the WIMP configuration becomes "fair"? The distinction between "technology" and "hardware" needs further clarification, to put it mildly. I suppose one should consider the cost. Why spend more for proprietary systems and hardware (not "technology"!) for worse performance?
FL
On 6/13/05, Michael Rhoadarmer Michael.R.Rhoadarmer@wheaton.edu wrote:
It really shouldn't surprise anyone that Open Source software
has issues
on a Windows box, especially 2000 which is running on a 6 year
old design.
The question might be would the same issue be there if you were running MSSQL and .NET instead of PHP and MySQL? Also, 512 is a lot of
RAM for LAMP
but its only a light snack for a Windows server. That's our
minimum standard
for office workstations. Still, for a fairer comparison of
technology (not
hardware) I would double the RAM, install 2003, and eliminate all non-essential services. I also might try it turning of page
files and going
straight RAM.
Why is building an even bigger server than the one already twice as fast a "fairer" comparison? How much does the LAMP configuration have to be handicapped before the WIMP configuration becomes "fair"? The distinction between "technology" and "hardware" needs further clarification, to put it mildly. I suppose one should consider the cost. Why spend more for proprietary systems and hardware (not "technology"!) for worse performance?
FL
Before this topic goes into OS war territory, let me clarify my motivation for the original post.
In times past, I've encountered PHP application performance differences between platforms, but never anything this noteworthy. Granted, MediaWiki is the largest 3rd party PHP app I've deployed, and I've not had the time to dig under the hood.
I just want to ensure I've not overlooked something in IIS or PHP which would improve the performance. Has anyone compared Windows/IIS vs Linux/Apache and found vast differences in MediaWiki performance?
----- Troy Thompson IT Manager, Envisionware, Inc tthompson@envisionware.com mailto:tthompson@envisionware.com 678-584-5911 x 212
mediawiki-l@lists.wikimedia.org