If you look a below code you will see measuring of performance of very simple for loop.

var res = 0;
function runs(){
    var a1 = performance.now();
    var x = 0;
    for(var i=0;i<10**9;i++) {
        x++;  
    }
    var a2 = performance.now();
    res += (a2-a1);
}

for(var j=0;j<10;j++){
    runs();
}
console.log(`=${res/10}`);

Additionally, just for a good measure, this will run 10 times and average results. Now, issue with this is that it is not reliable, it highly depends on your CPU, memory and other programs running on your device. First time, may run 9s and second 23s, then subsequent call can result 8s.

Is there a way to measure performance regardless of CPU, memory and everything else? I am after something that will give relative number of FLOPS or any other measure that when you compare two codes you will exactly know that one code executes faster than the other.

For instance for loop with 1005 will always show slower than one with 1000 iterations.

Note: Saying FLOPS is wrong in this context as it means - floating point operations per second. I would like to exclude time completely, I only need FPOs. Meaning I do not care about seconds, just about reliably knowing that regardless of device if you have same code it will always take let say 2000 FPOs to execute.