Fahrvergnuugen
I am the law!
Its been said before in many benchmarks that I've read that macs often make better use of dual processors than dual processor PCs running windows (not sure about *nix).
Well, I was running some benchmarks of my own using PHP which is not MP aware. These benchmarks are being used as ammo to get the company I work for to let me have an xServe or at least linux / apache as the webserver that I write code for.
I wrote a processor intensive XML parsing PHP script. The XML file is only 18MB so disk speed and ram aren't of a major consequence to the execution of the script.
here is where it gets interesting
On a dual processor windows 2000 server, the script will pin one processor at 100% leaving the other processor untouched.
On my Dual processor G5, php will use ~50% of BOTH processors according to top.
Strange eh?
If you're interested in the results, here they are:
System 1 (steelmagnolia)
Dual 2.0 GHz G5 running OSX 10.3 / Apache
22.5 seconds
System 2 (acc-intranet)
Dual 1.4 GHz Pentium 3 running Windows 2000 Server / IIS
55.0 seconds
System 3 (acc-webdev)
Single 1.4 GHz Pentium 3 running Windows 2000 Server / IIS
57.8 seconds
System 4
Single 1.26 GHz Pentium 3 running Slackware Linux 8.0 / Apache
37.8 seconds
System 5
Dual 500MHz G4 running OSX 10.3 / apache
90 seconds (exactly 4 times the dual 2GHz G5)
You'll notice that the slower 1.26GHz P3 with linux was 31% faster than the 1.4GHz P3 with windows.
Well, I was running some benchmarks of my own using PHP which is not MP aware. These benchmarks are being used as ammo to get the company I work for to let me have an xServe or at least linux / apache as the webserver that I write code for.
I wrote a processor intensive XML parsing PHP script. The XML file is only 18MB so disk speed and ram aren't of a major consequence to the execution of the script.
here is where it gets interesting
On a dual processor windows 2000 server, the script will pin one processor at 100% leaving the other processor untouched.
On my Dual processor G5, php will use ~50% of BOTH processors according to top.
Strange eh?
If you're interested in the results, here they are:
System 1 (steelmagnolia)
Dual 2.0 GHz G5 running OSX 10.3 / Apache
22.5 seconds
System 2 (acc-intranet)
Dual 1.4 GHz Pentium 3 running Windows 2000 Server / IIS
55.0 seconds
System 3 (acc-webdev)
Single 1.4 GHz Pentium 3 running Windows 2000 Server / IIS
57.8 seconds
System 4
Single 1.26 GHz Pentium 3 running Slackware Linux 8.0 / Apache
37.8 seconds
System 5
Dual 500MHz G4 running OSX 10.3 / apache
90 seconds (exactly 4 times the dual 2GHz G5)
You'll notice that the slower 1.26GHz P3 with linux was 31% faster than the 1.4GHz P3 with windows.
