为已知的更常见路径优化分支
请考虑以下代码:
void error_handling();
bool method_impl();
bool method()
{
const bool res = method_impl();
if (res == false) {
error_handling();
return false;
}
return true;
}
我知道 method_impl()
会返回 true
99.999%(是的,三位小数),但我的编译器不会.method()
在时间消耗方面是部分关键的.
I know method_impl()
will return true
99.999% (yes, three decimal places) of the time, but my compiler doesn't. method()
is partially critical in term of time-consumption.
- 我是否应该重写
method()
(并降低可读性)以确保仅在method_impl()
返回false
时才可能发生跳转?如果是,如何? - 我应该让编译器为我做这些工作吗?
- 我应该让 CPU 的分支预测为我做这些工作吗?
- Should I rewrite
method()
(and make it less readable) to ensure a jump may only occur whenmethod_impl()
returnsfalse
? If yes, how? - Should I let the compiler do the work for me?
- Should I let the branch prediction of my CPU do the work for me?
推荐答案
底层硬件已经执行了这种优化.第一次预测它会失败",但在它会命中正确的选项 en.wikipedia.org/wiki/Branch_predictor 之后.
The underlying hardware already performs this optimizations. It will "fail" to predict it the first times, but after it will hit the correct option en.wikipedia.org/wiki/Branch_predictor.
您可以尝试应用 GCC 扩展并检查使用它是否更快,但我认为您几乎看不出有和没有它的任何区别.总是应用分支预测,它不是你启用的东西
You can try applying the GCC extension and check if it is faster with it or not, but I think you will barely see any difference with it and without it. The branch prediction is applied always, it is not something you enable
相关文章