Why running benchmark tests in the console gives different results

I wanted to test the effectiveness of various ways of checking if a value is an array. I've used console.time and jsperf and got conflicting results. Generating the script and running it gave similar jsperf results.

jsperf link http://jsperf.com/is-it-an-array

my code

a={test:'a',foo:'bar'};
b='abcdef';
c=[1,2,3,4,5,6]
d=3.14

length=1e6;
l=length;
console.time(1);
while(l--){
  a instanceof Array ?true:false;
  b instanceof Array ?true:false;
  c instanceof Array ?true:false;
  d instanceof Array ?true:false;
}
console.timeEnd(1);
l=length;
console.time(2);
while(l--){
  a.pop?true:false;
  b.pop?true:false;
  c.pop?true:false;
  d.pop?true:false;
}
console.timeEnd(2);
l=length;
console.time(3);
while(l--){
  Array.isArray(a)?true:false;
  Array.isArray(b)?true:false;
  Array.isArray(c)?true:false;
  Array.isArray(d)?true:false;
}
console.timeEnd(3);
l=length;
console.time(4);
while(l--){
  a.constructor == Array?true:false;
  b.constructor == Array?true:false;
  c.constructor == Array?true:false;
  d.constructor == Array?true:false;
}
console.timeEnd(4);

      

        
        
        
      

    

On my computer Firefox works better with them than Chrome using jsperf. The fastest solution is to use instanceof

and the worst is .pop

. In chrome it isArray

is the fastest and .pop

still the slowest.

But running these tests in the console .pop

will never get slower.

Firefox

1: 153.35ms
2: 166.88ms //.pop
3: 169.94ms
4: 193.73ms

      

        
        
        
      

    

Chromium

1: 286.000ms
2: 207.000ms //.pop
3: 270.000ms
4: 266.000ms

      

        
        
        
      

    

What does using the console to run code give different results?