性能

60帧每秒，16000微秒每帧。

function perf(name, exec) {
    var N = 100000, i;
    var idle = function(){};
    var start = getTimer();
    i = N; while (i--) idle();
    var idleRate = (getTimer() - start) * 1000 / N;
    start = getTimer();
    i = N; while (i--) exec();
    var rate = (getTimer() - start) * 1000 / N;
    trace(name, Math.round((rate - idleRate) * 1000) / 1000, 'us/op');
}

循环方法

while循环最好。

N = 10000000;
start = getTimer();
for (var i = 0; i < N; i++);
trace('for-loop', (getTimer() - start) * 1000 / N, 'us/op');

start = getTimer();
for (var i = N; i--;);
trace('for-loop2', (getTimer() - start) * 1000 / N, 'us/op');

start = getTimer();
var i = N;
while (i--);
trace('while-loop', (getTimer() - start) * 1000 / N, 'us/op');

//for-loop 1.2013 us/op
//for-loop2 0.7189 us/op
//while-loop 0.6223 us/op

Utils.rgb()

手动比Utils.rgb()快一倍。主要时间花在函数调用。

rgb = Utils.rgb;
r, g, b, result;

perf('manual', function() { r << 16 | g << 8 | b; });
perf('Utils.rgb()', function() { rgb(r, g, b); });

//manual 1.488 us/op
//Utils.rgb() 2.998 us/op

数组访问方法

Array.[]略快于Vector.[]。（.forEach()和foreach()慢较多，结果未写在这里，时间花在函数调用。）

a = [1];
v = $.toIntVector(a);

perf('Array.[]', function(){ a[0]; });
perf('Vector.[]', function(){ v[0]; });

//Array.[0] 0.676 us/op
//Vector.[] 0.797 us/op

时间精度初测

初测getTimer()获取的精度略高于Player.time。输出端似乎只有interval。方法二整合二者

clear();
var bi = 0;
var _i = getTimer();
var _is = Player.time;
interval(function(){
trace("时间("+5 * (bi++)+"ms):");
trace("Player.time:"+Player.time);
trace("方法二："+(_is+(getTimer()-_i)));
},5,50);