Performance tests I done E3-1240 VS E3-1230 VS L5520 VS Atom 230

[inklight]

Hi all ,
I don't know if this the right section i hope it the closest .

I have many VPS server uses versus CPU hardware that give me chance to some of my own tests .
this tests not relative directly to passmark softwares but could give Idea about real performance of those CPu's
Test tool I use it
www.php-benchmark-script.com

This is my PHP benchmark Results

----------------------------------------------------------------------------
VPS Server : E3-1240 V3 3.4GHz PassmarkMark-9731 1Core Single 2199
--------------------------------------
test_math : 1.362 sec.
test_stringmanipulation : 1.318 sec.
test_loops : 0.688 sec.
test_ifelse : 0.599 sec.
--------------------------------------
Total time: : 3.9 sec.
--------------------------------------
VPS Server : E3-1240 V2 3.40GHz PassmarkMark-9302 1Core Single 2034
test_math : 1.477 sec.
test_stringmanipulation : 1.641 sec.
test_loops : 0.855 sec.
test_ifelse : 0.590 sec.
--------------------------------------
Total time: : 4.5 sec.
--------------------------------------
VPS Server : E3-1230 3.30GHz PassmarkMark-8839 1Core Single 1932
--------------------------------------
test_math : 2.123 sec.
test_stringmanipulation : 2.041 sec.
test_loops : 0.960 sec.
test_ifelse : 0.711 sec.
--------------------------------------
Total time: : 5.8 sec.
--------------------------------------
VPS Server : 2X L5520 2.27GHz PassmarkMark-7156 1Core Single 1034
--------------------------------------
test_math : 3.438 sec.
test_stringmanipulation : 3.058 sec.
test_loops : 1.368 sec.
test_ifelse : 0.986 sec.
--------------------------------------
Total time: : 8.85 sec.
----------------------------------------------------------------------------
Dedi Server : Atom 230 1.60GHz Mark-295 1Core Single 242
test_math : 11.173 sec.
test_stringmanipulation : 10.948 sec.
test_loops : 5.985 sec.
test_ifelse : 4.017 sec.
--------------------------------------
Total time: : 32.1 sec.
--------------------------------------

I notice Huge different between Low end Process like Atom 230 and single core of Xeon Process Reach 7X faster .
I hope test help some before take decision bu/rent any of this CPu's

[David (PassMark)]

The PHP benchmark isn't really representative of the hardware performance because the PHP script will only be single threaded. Which is why the dual CPU L5520 machine didn't do so well.

Different versions of PHP are also likely to give different results. The script also doesn't use much RAM, and no disk activity, so isn't all that realistic of what a typical PHP script would be doing.

Of course if the PHP script was threaded the Atom would look even worse than it already does.

[inklight]

The PHP benchmark isn't really representative of the hardware performance because the PHP script will only be single threaded. Which is why the dual CPU L5520 machine didn't do so well.
Of course if the PHP script was threaded the Atom would look even worse than it already does.

Thanks for valuable information . The L5520 was one core from one CPU the same with others VPS they all one core only .
My main concern is with PHP because I build sites and was looking for best optimization .
whats make me doing this test is poor performance Atom 230 sever on site on it . comparing to other VPS's . Specially I use same script (Wordpress) Same PHP version 5.4 and same APCu version .
when you said single threaded do you mean php will use one core no matter what . so Dedi E3-1240 sever will give same performance of E3-1240 one core VPS .
BTW from my all test I notice PHP makes the biggest Bottleneck on my servers for example MySQL do queries in milliseconds .

[David (PassMark)]

you said single threaded do you mean php will use one core no matter what

That benchmark script is single threaded. Meaning it will only use 1 CPU core.

But in real life a web site can have many simultaneous users. So more than 1 core could be used, depending on the site.

We did a comparison between PHP and other scripting options, including native code.
http://www.wrensoft.com/zoom/benchmarks.html
Native code can be 5x to 10x faster than PHP.