Mail Archives: cygwin/2004/05/27/10:55:06
Andrew DeFaria wrote:
> My understanding is that the Cygwin port of Apache 1.x is also
> significantly slower than the native Apache 1.x but this didn't stop
> people from wanting a Cygwin version of 1.x. Or is there something in
> 2.x (this thread MPM thing) that would make a Cygwin port of Apache 2.x
> much, much slower than the native one?
You're correct that 1.x also suffers a performance penalty compared to
the native win32 version. However 1.x can only operate in the prefork
mode which is not suitable to Windows since process creation is
relatively expensive. Thus both versions are pokey. However, with the
advent of 2.x the method of allocating workers is modular (the MPM) and
so you can choose to have them as threads or as the old prefork style,
among others. With threads the performance under Windows is much
improved. I suspect (but have not tested) that the Cygwin overhead
would be even more apparent in that case, because 2.x has been
specifically designed to get good performance under win32 natively,
whereas 1.x was never intended for such systems.
As far as I know the popularity of Cygwin Apache is for developing and
testing web applications that will eventually reside on unix servers.
In that department 1.x is more popular by a huge margin due to its
stability, known quirks, and ability to work well with non-thread-safe
PHP extensions (as well as general stubbornness of sysadmins who avoid
2.x.) Thus demand for 1.x/Cygwin should naturally be much higher as
well. If you're actually interested in running a server then the native
version is probably better.
> BTW: Thanks for volunteering for this. Does this mean that a Cygwin
> version of mod_php would be working again?
Yes, I intend to do that.
Brian
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -