Re: APC's performance improvement patch for single process multithreaded php
| From: | Basant Kukreja | Date: | Sun, 11 Oct 2009 04:28:45 +0000 |
| Subject: | Re: APC's performance improvement patch for single process multithreaded php | ||
| References: | 1 2 3 4 | Groups: | php.pecl.dev |
| Request: | Send a blank email to pecl-dev+get-6688@lists.php.net to get a copy of this message | ||
> You can't set apc.stat to 30 seconds, it's a binary flag so it's either on
> or off. In the off scenario this isn't an issue as the files will never be
> updated unless you happen to cache two different versions on the first
> request for each process.
Agreed. I have used it apc.stat many times and confused with interval.
Sorry about that.
I now agree with you that anonymous map and heap allocation will perform
identical.
I measure it on an e-commerce workload with and without this patch
on Solaris Sparc and found no difference between the two.
Excl. Incl. Excl. Incl. Excl. Incl. Name
User CPU User CPU Total LWP Total LWP Sys. CPU Sys. CPU
sec. sec. sec. sec. sec. sec.
Using calloc:
4.243 7656.376 7.145 56560.845 0.610 2095.696 php5_execute
Using mmap:
3.973 7644.517 6.715 54048.538 0.640 2071.839 php5_execute
I think I spoke too early without measurement.
I will close the RFE.
Thanks for your input.
Regards,
Basant.