Re: Re: svn: /php/php-src/ branches/PHP_5_3/ext/date/php_date.c trunk/ext/date/php_date.c

On Tue, Feb 1, 2011 at 10:23 AM, Stas Malyshev <smalyshev@sugarcrm.com> wrote: > Though, of course, maybe I'm wrong - what would you suggest is the problem > then? I saw similar slowdowns on 1-process benchmarks as on multi-process > server benchmarks - so I have hard time believing it has anything to do with > threading vs. processes. See what we do on each path in TS mode and what we do in NTS, and that's what I meant by tons of useless repetitive ops. >> more clear, I'm not (hell not) talking abut parallelism, or sharing >> engine's data across threads but how we achieve thread safety for the >> shared data or which data we have to share or not, or TS APIs (*_r). > > So what specifically do you think can be done better there? TLS is one part and then identify the part of the engine/exts that need TS and use better locking mechanism (rwlock-like). But my main point is, for windows, that process based application will never perform as well as they could with threads. And that's a fact, every windows developer knows that. But this discussion is a bit off topic :) Cheers, -- Pierre @pierrejoye | http://blog.thepimp.net | http://www.libgd.org