I also tested a few different configs on test.php
edit: seems I have apache keepalives off and apachebench was run without keepalives, so will have to redo some tests
- apache standalone
- apache + varnish 2.1.4 defaults on port 8888
- litespeed no cache + varnish 2.1.4 defaults on port 8888
just post summary here
system is
Quote:
|
System is virtualbox virtual guest CentOS 5.5 64bit install with Xeon W3540 @3408Mhz, 640GB Samsung SATAII drive OS disk with 1GB out of 6GB DDR3-1550Mhz 9-9-9-24 memory speed assigned to guest OS.
|
Code:
ab -n 1000 -c 200 192.168.56.101/test.php
apache standalone
Code:
1st run - Requests per second: 4496.60
2nd run - Requests per second: 1309.79
3rd run - Requests per second: 408.32
4th run - Requests per second: 120.17
apache + varnish 2.1.4 defaults on port 8888
Code:
1st run - Requests per second: 4240.09 (uncached)
2nd run - Requests per second: 9565.99
3rd run - Requests per second: 8947.67
4th run - Requests per second: 8688.70
5th run - Requests per second: 8600.38
average = 8,008 rps (including uncached 1st response) or 8,950 rps varnish cached
litespeed no cache + varnish 2.1.4 defaults on port 8888
Code:
1st run - Requests per second: 10661.32
2nd run - Requests per second: 9079.93
3rd run - Requests per second: 8776.55
4th run - Requests per second: 8805.22
5th run - Requests per second: 9213.36
average = 9,307 rps
Looks like Litespeed cache still leading for small test.php file at least at 13-20K rps - nearly 2-5x times faster than litespeed no cache, or up to 2x times faster than apache + varnish defaults or litespeed + varnish defaults