Re: [apache-plusplus] Something else to think about...

[bret <mersault@busboy.sped.ukans.edu>]

> > 	Wby not go with the idea of a single thread doing everything... basically
> > replacing the processes in the current Apache implementation with threads?
> 
> That's what we're already doing... there's at least four threaded versions
> of apache now... two aren't publically available, one is the NT version of
> apache 1.3, and the other is apache-nspr.
> 
> > Okay, so that's a simplification, but you get the point... we slap a mutex
> > around accept(), which is performed within each thread, thereby removing
> > the need for a _single_ condition variable, and watch performance
> > skyrocket.
> 
> On linux, the performance improves, but doesn't skyrocket.  I'm talking
> maybe a 10% improvement in crude benchmarks.  There's still more to be
> gained, but it's not going to be a huge change.

	Okay, "huge gain" was hyperbole!  And the more I thought about it,
the more I thought that a single reader/multithreaded handler might be
better for handling pipelined requests, even if it is just a touch slower
than optimum.  And we are interested in correctness over performance
here...
	Your thoughts, Dean?  Or anybody else, for that matter?


								- Bret -