FluffyCat.com . PHP Design Patterns PHP Performance Tuning if VS switch

PHP Design Patterns PHP Performance Tuning if VS switch

PHP Performance Tuning - if /elseif VS switch case


I often use the php switch case statement because I think it looks neater in the code and because I think it is probably more efficient to run. But is it more efficient? I set up this little test to run if / elseif VS switch case 100,000 times each.

The results - if / else was slightly better. While I still think switch / case looks neater it is actually slower. If you are really squeezing all the speed you can get from your application if / elseif may be the way to go.




testPerformanceIfVSSwitch

//copyright Lawrence Truett and FluffyCat.com 2009, all rights reserved
  	define('BR', '<'.'BR'.'>');
	echo 'Performance Testing PHP if / else VS switch with 100,000 iterations'.BR.BR;
	
	$ifTime = 0;
	$switchTime = 0;
	for ($x = 0; $x < 100000; $x++) {
		$oneToFive = rand(1,5);
		if (0 == fmod($x,2)) {
			$switchTime = $switchTime + testSwitch($oneToFive);
			$ifTime = $ifTime + testIf($oneToFive);
		} else {
			$ifTime = $ifTime + testIf($oneToFive);  			
			$switchTime = $switchTime + testSwitch($oneToFive);  		
		}
	}
	echo 'total if / else time: '.$ifTime.BR;
	echo 'total case time: '.$switchTime.BR;
	echo BR;
	if ($ifTime > $switchTime) {
		echo 'switch is quicker by '.($ifTime - $switchTime).BR;
	} else {
		echo 'if / else is quicker by '.($switchTime - $ifTime).BR;		
	}


	function testSwitch($oneToFive) {
		$time_start = microtime(true);
		switch($oneToFive) {
			case 1: $z = 1; break;
			case 2: $z = 2; break;
			case 3: $z = 3; break;
			case 4: $z = 4; break;
			case 5: $z = 5; break;						
		}
		$time_end = microtime(true);
		$time = $time_end - $time_start;
		return $time;
	}
  
	function testIf($oneToFive) {
		$time_start = microtime(true);
		if (1 == $oneToFive) {
			$z = 1;
		} elseif(2 == $oneToFive) {
			$z = 2;			
		} elseif(3 == $oneToFive) {
			$z = 3;	
		} elseif(4 == $oneToFive) {
			$z = 4;	
		} elseif(5 == $oneToFive) {
			$z = 5;	
		}
		$time_end = microtime(true);
		$time = $time_end - $time_start;
		return $time;
	}
download source, use right-click and "Save Target As..." to save with a .php extension.


output 1 of testPerformanceIfVSSwitch.php

Performance Testing PHP if / else VS switch with 100,000 iterations


total if / else time: 0.66318321228027
total case time: 0.68480849266052


if / else is quicker by 0.021625280380249


output 2 of testPerformanceIfVSSwitch.php

Performance Testing PHP if / else VS switch with 100,000 iterations


total if / else time: 0.66332268714905
total case time: 0.68789744377136


if / else is quicker by 0.024574756622314


output 3 of testPerformanceIfVSSwitch.php

Performance Testing PHP if / else VS switch with 100,000 iterations


total if / else time: 0.6631076335907
total case time: 0.67988753318787


if / else is quicker by 0.016779899597168


output 4 of testPerformanceIfVSSwitch.php

Performance Testing PHP if / else VS switch with 100,000 iterations


total if / else time: 0.66344714164734
total case time: 0.67861771583557


if / else is quicker by 0.015170574188232


output 5 of testPerformanceIfVSSwitch.php

Performance Testing PHP if / else VS switch with 100,000 iterations


total if / else time: 0.66696906089783
total case time: 0.68225598335266


if / else is quicker by 0.015286922454834
Comments Comments are left by visitors to FluffyCat.com, are not endorsed by FluffyCat.com, and may or may not be accurate.
Comment by spicerje Rate this Comment

Orginal only 5
Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.36938524246216
total case time: 0.37775444984436

if / else is quicker by 0.0083692073822021

I increased the code to 10 and the if /else still wins
Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.38648891448975
total case time: 0.39371657371521

if / else is quicker by 0.0072276592254639

I then increase to 20 and still if/else wins

Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.41207385063171
total case time: 0.41935849189758

if / else is quicker by 0.0072846412658691

I finally increased the code to 50 and now the switch statement beats the if /else

Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.55141520500183
total case time: 0.49664068222046

switch is quicker by 0.054774522781372

so for most uses it appears that the if/else is faster, but the
switch looks nicer. For large amounts of comparisons
the switch will be faster.

Comment by krishnabhaya Rate this Comment

HI,it's wonder that same script, showing me switch case is quicker in all cases.

I got
Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.64044070243835
total case time: 0.59019613265991

switch is quicker by 0.050244569778442
--------------
Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.81924557685852
total case time: 0.61534929275513

switch is quicker by 0.20389628410339

so there must be some other factor, which affects the same.

Comment by Larry Rate this Comment

I wondered what increasing the number of if / case statements would do to the results. Increasing from 5 to 10 didn't change the results that much. The total times increased only slightly, and the if / elseif was still faster - although by a smaller margin.

Sample result with 10 choices:

Performance Testing PHP if / else VS switch with 100,000 iterations

total if / else time: 0.69219636917114
total case time: 0.70317435264587

if / else is quicker by 0.010977983474731

Sign In
to add your own comment