ASP.NET, PHP, ASP Performance Matrix

I was reading an article online comparing PHP to ASP.NET the other day. The author, an Oracle employee, had some valid claims regarding PHP’s advantages when it comes to costs and flexibility, but he also had some unfounded claims that PHP was more efficient and faster. I beg to differ, so I ran some tests using a similar function for all three languages.

Test Platform

The test was run on an AMD Athlon XP 2800+ with 2 GB of ram running Windows XP Professional SP2. We used a standard install of Internet Information Services and gave PHP the added benefit of installing Zend Optimizer v2.6.2. We used the php5isapi.dll to map and execute our PHP scripts.

• AMD Athlon XP 2800+ (2.08 GHz)
• 2 GB Ram
• Windows XP Professional SP2
• Internet Information Services v5.1
• Zend Optimizer v2.6.2
• PHP 5.1.2
• ASP 3.0

Test Methodology

We wrote a similar script to test each language. It was a simple for loop that was programmed to loop 80 million times. You might need to increase the max execution settings in the php.ini file for PHP, or increase it programmatically. We chose to do it in the php.ini to make sure it was set correctly. The same holds true for ASP. You may need to increase the Server.ScriptTimeout value.

Some lines below might wrap, so be sure to check that before attempting to execute the code.

ASP Code

<%

'Server.ScriptTimeout = 360


'Response.Buffer = False





Dim i

Dim startTime


Dim endTime

Dim totalTime



For x = 1 To 10


  startTime = Now

  Response.Write("Start Time: " & startTime & "<br />")

 

  For i = 1 To 80000000

    If i > 0 Then

    End If

  Next



  endTime = Now


  totalTime = endTime - startTime





  Response.Write("End Time: " & endTime & "<br />")


  Response.Write("Total time: " & Minute(totalTime) & ":" & Second(totalTime) & "<br /><br />")

Next


%>

PHP Code

<?php

function getmicrotime()

{

  list($usec, $sec) = explode(" ",microtime());

  return ((float)$usec + (float)$sec);

}



for ($x=0; $x < 10; $x++)

{


  $number = 0;

  $time_start = getmicrotime();



  for ($i=0; $i < 80000000; $i++)

  {

    if ("hello" == "world") { }

  }



  $time_end = getmicrotime();


  $time = $time_end - $time_start;



  echo "Did 80M
additions in $time seconds";

}

?>

ASP.NET (VB.NET) Code

<%@ Language="VB" %>

<script runat="server" language="vb">

Private Sub Page_Load(ByVal sender As System.Object, ByVal e As System.EventArgs)

  Dim i As Long

  Dim x As Int16

  Dim startTime As Decimal

  Dim endTime As Decimal

  Dim totalTime As Decimal



  For x = 1 To 10

    startTime = Convert.ToDecimal(DateTime.Now.ToString("ss.fff"))

    Response.Write("Start Time: " & startTime.ToString() & "<br />")

    

    For i = 1 To 80000000

      If "hello" Is "world" Then

      End If

    Next



    endTime = Convert.ToDecimal(DateTime.Now.ToString("ss.fff"))

    totalTime = endTime - startTime



    Response.Write("End Time: " & endTime.ToString() & "<br />")

    Response.Write("Total time: " & totalTime.ToString() & "<br /><br />")

  Next

End Sub

Results

It’s no surprise that ASP.NET is the clear winner. The results that did surprise me were that Zend Optimizer consistently increased PHP’s performance by nearly 50%. That is a huge number. It is nonsense not to use PHP with the Optimizer according to our study.

The PHP results were roughly double of the numbers listed below without the Zend Optimizer running. And, as you can see, ASP’s performance is just plain bad.

ASP.NET	PHP	ASP
0.14	15.58716	33
0.125	15.92204	33
0.157	15.58681	35
0.124	15.58992	35
0.125	15.59410	33
0.125	15.93063	34
0.141	15.60623	33
0.125	15.60107	33
0.14	15.60247	34
0.125	16.14546	35

Results are in seconds

We welcome and performance enhancements and/or your own results. Just post your thoughts, results and any suggestions. I want to extend special thanks to Rufus Harvey for his technical prowess in helping put this test together!

- Ernest Carroll