Hi Thies,
I'll see which parts of your patch can be adapted for 5.1.x. I think the var lookup is probably the most substantial one but I will probably implement it slightly different. I'll mail internals@ when I have something.
As I mailed you, I tried the zval_copy_ctor() and zval_dtor() optimization of yours but didn't get any improvement. Can you try and run the patch I sent you to see if you get different results?
I'm also attaching it again, in case others want to see if it changes anything.
Andi
At 06:36 PM 9/25/2004 +0200, Thies C. Arntzen wrote:
On 2004-09-24 01:42:48 +0200, Andi Gutmans <[EMAIL PROTECTED]> said:
Hi,
Take it at face value but I get the following results with the new VM vs. 5.0.x on the bench.php script I commited to php5/Zend/tests.
PHP 5.0.x PHP 5.1.x
37.5secs vs 29secs
You know I'm not a big believer in synthetic benchmarks but it's still quite nice.
Andi
on my box:
5.0.x 45,3 5.1.x 36.1 4.3.x(*) 27.1
i just could not resist...
re, tc
*with fasterharderstronger patch
-- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
-- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
