Re: [PHP4BETA] RFC: Request startup optimization

I also remember having liked this idea in the beginning. But the more we talked about it at the PDM the more I saw it won't really help performance too much and could actually degrade performance. Andi At 08:16 PM 2/11/00 +0200, Zeev Suraski wrote:

On Fri, 11 Feb 2000 thies@digicol.de wrote: On Fri, Feb 11, 2000 at 07:05:20PM +0100, Sascha Schumann wrote:

    my initial "hello world" benchmark showed no measurable jump
    at all (could be me beeing stupid - will do some quantifying
    early next week).

    That's what I almost expected. How many modules did your test
    installation contain?

    obviously not enough (one or two) - so the results could be
    counted as "expected".

Even if you don't compile any modules into PHP, it includs at least a dozen built in modules... The truth is that the startup functions really pose no noticeable overhead, even if you were to completely remove them. If we had a startup function that actually calculates something, or does something more complex than a couple of assignments and a single emalloc(), it would have been worth it. But I can't think of any such request startup function. Zeev

--- Andi Gutmans <andi@zend.com> http://www.zend.com/