内联例程(消除调用/返回和参数推送) 试着用表查找(如果它们更快的话)消除测试/开关 展开循环(Duff的设备)到刚好适合CPU缓存的位置 本地化内存访问，以免耗尽缓存 如果优化器还没有本地化相关的计算 如果优化器还没有这样做，就消除循环不变量

我大半辈子都在这里度过。大致的方法是运行你的分析器并记录它:

Cache misses. Data cache is the #1 source of stalls in most programs. Improve cache hit rate by reorganizing offending data structures to have better locality; pack structures and numerical types down to eliminate wasted bytes (and therefore wasted cache fetches); prefetch data wherever possible to reduce stalls. Load-hit-stores. Compiler assumptions about pointer aliasing, and cases where data is moved between disconnected register sets via memory, can cause a certain pathological behavior that causes the entire CPU pipeline to clear on a load op. Find places where floats, vectors, and ints are being cast to one another and eliminate them. Use __restrict liberally to promise the compiler about aliasing. Microcoded operations. Most processors have some operations that cannot be pipelined, but instead run a tiny subroutine stored in ROM. Examples on the PowerPC are integer multiply, divide, and shift-by-variable-amount. The problem is that the entire pipeline stops dead while this operation is executing. Try to eliminate use of these operations or at least break them down into their constituent pipelined ops so you can get the benefit of superscalar dispatch on whatever the rest of your program is doing. Branch mispredicts. These too empty the pipeline. Find cases where the CPU is spending a lot of time refilling the pipe after a branch, and use branch hinting if available to get it to predict correctly more often. Or better yet, replace branches with conditional-moves wherever possible, especially after floating point operations because their pipe is usually deeper and reading the condition flags after fcmp can cause a stall. Sequential floating-point ops. Make these SIMD.

我还喜欢做一件事:

将编译器设置为输出程序集清单，并查看它为代码中的热点函数发出了什么。所有那些聪明的优化，“一个好的编译器应该能够自动为你做”?实际的编译器可能不会执行这些操作。我见过GCC发出真正的WTF代码。

When you get to the point that you're using efficient algorithms its a question of what you need more speed or memory. Use caching to "pay" in memory for more speed or use calculations to reduce the memory footprint. If possible (and more cost effective) throw hardware at the problem - faster CPU, more memory or HD could solve the problem faster then trying to code it. Use parallelization if possible - run part of the code on multiple threads. Use the right tool for the job. some programing languages create more efficient code, using managed code (i.e. Java/.NET) speed up development but native programing languages creates faster running code. Micro optimize. Only were applicable you can use optimized assembly to speed small pieces of code, using SSE/vector optimizations in the right places can greatly increase performance.

OK, you're defining the problem to where it would seem there is not much room for improvement. That is fairly rare, in my experience. I tried to explain this in a Dr. Dobbs article in November 1993, by starting from a conventionally well-designed non-trivial program with no obvious waste and taking it through a series of optimizations until its wall-clock time was reduced from 48 seconds to 1.1 seconds, and the source code size was reduced by a factor of 4. My diagnostic tool was this. The sequence of changes was this:

The first problem found was use of list clusters (now called "iterators" and "container classes") accounting for over half the time. Those were replaced with fairly simple code, bringing the time down to 20 seconds. Now the largest time-taker is more list-building. As a percentage, it was not so big before, but now it is because the bigger problem was removed. I find a way to speed it up, and the time drops to 17 seconds. Now it is harder to find obvious culprits, but there are a few smaller ones that I can do something about, and the time drops to 13 sec.

现在我似乎遇到了瓶颈。样本告诉我它到底在做什么，但我似乎找不到任何可以改进的地方。然后，我考虑了程序的基本设计及其事务驱动结构，并询问它所做的所有列表搜索实际上是否都是由问题的需求强制执行的。

然后我偶然发现了一种重新设计，在这种设计中，程序代码实际上是从一组较小的源代码中生成的(通过预处理器宏)，在这种设计中，程序不会不断地找出程序员知道的相当可预测的事情。换句话说，不要“解释”要做的事情的顺序，要“编译”它。

重新设计完成了，源代码缩减了1 / 4，时间减少到10秒。

现在，因为它变得如此之快，很难进行抽样，所以我给它10倍的工作，但下面的时间是基于原始工作负载的。

进一步的诊断表明，它是在队列管理上花费时间的。内联这些将时间缩短到7秒。 现在一个很大的时间消耗是我一直在做的诊断打印。冲水- 4秒 现在最浪费时间的是调用malloc和free。回收对象- 2.6秒。 继续进行抽样，我仍然发现了严格意义上没有必要的操作——1.1秒。

总加速系数:43.6

Now no two programs are alike, but in non-toy software I've always seen a progression like this. First you get the easy stuff, and then the more difficult, until you get to a point of diminishing returns. Then the insight you gain may well lead to a redesign, starting a new round of speedups, until you again hit diminishing returns. Now this is the point at which it might make sense to wonder whether ++i or i++ or for(;;) or while(1) are faster: the kinds of questions I see so often on Stack Overflow.

附注:可能有人想知道我为什么不用侧写器。答案是，几乎所有这些“问题”都是函数调用站点，堆栈样本可以精确定位。即使在今天，分析人员也只是勉强接受这样一个观点:语句和调用指令比整个函数更重要，更容易定位，也更容易修复。

我实际上构建了一个剖析器来做这件事，但是要真正了解代码正在做什么，没有什么可以替代您的手指。样本数量少并不是问题，因为被发现的问题没有一个小到容易被忽略的程度。

添加:jerryjvl要求一些例子。这是第一个问题。它由少量独立的代码行组成，加在一起占用了一半的时间:

 /* IF ALL TASKS DONE, SEND ITC_ACKOP, AND DELETE OP */
if (ptop->current_task >= ILST_LENGTH(ptop->tasklist){
. . .
/* FOR EACH OPERATION REQUEST */
for ( ptop = ILST_FIRST(oplist); ptop != NULL; ptop = ILST_NEXT(oplist, ptop)){
. . .
/* GET CURRENT TASK */
ptask = ILST_NTH(ptop->tasklist, ptop->current_task)

These were using the list cluster ILST (similar to a list class). They are implemented in the usual way, with "information hiding" meaning that the users of the class were not supposed to have to care how they were implemented. When these lines were written (out of roughly 800 lines of code) thought was not given to the idea that these could be a "bottleneck" (I hate that word). They are simply the recommended way to do things. It is easy to say in hindsight that these should have been avoided, but in my experience all performance problems are like that. In general, it is good to try to avoid creating performance problems. It is even better to find and fix the ones that are created, even though they "should have been avoided" (in hindsight). I hope that gives a bit of the flavor.

下面是第二个问题，分两行:

 /* ADD TASK TO TASK LIST */
ILST_APPEND(ptop->tasklist, ptask)
. . .
/* ADD TRANSACTION TO TRANSACTION QUEUE */
ILST_APPEND(trnque, ptrn)

它们通过在列表的末尾附加项目来构建列表。(解决方法是将项目收集到数组中，并一次性构建列表。)有趣的是，这些语句只花费了原始时间的3/48(即在调用堆栈上)，所以它们实际上在一开始并不是一个大问题。然而，在消除了第一个问题后，它们只花费了3/20的时间，所以现在是一条“大鱼”。总的来说，就是这样。

我可以补充说，这个项目是从我参与的一个真实项目中提炼出来的。在那个项目中，性能问题要严重得多(加速也是如此)，比如在内部循环中调用数据库访问例程来查看任务是否完成。

参考补充道: 源代码，无论是原始的还是重新设计的，都可以在www.ddj.com上找到，1993年，文件9311.zip, files slug。Asc和slug.zip。

编辑2011/11/26: 现在有一个SourceForge项目包含了Visual c++中的源代码，以及它是如何调优的详细描述。它只经历了上述场景的前半部分，并不完全遵循相同的顺序，但仍然获得了2-3个数量级的加速。

内联例程(消除调用/返回和参数推送) 试着用表查找(如果它们更快的话)消除测试/开关 展开循环(Duff的设备)到刚好适合CPU缓存的位置 本地化内存访问，以免耗尽缓存 如果优化器还没有本地化相关的计算 如果优化器还没有这样做，就消除循环不变量

在带有模板的语言(c++ /D)中，您可以尝试通过模板参数传播常量值。你甚至可以用开关来处理小的非常值集合。

Foo(i, j); // i always in 0-4.

就变成了

switch(i)
{
    case 0: Foo<0>(j); break;
    case 1: Foo<1>(j); break;
    case 2: Foo<2>(j); break;
    case 3: Foo<3>(j); break;
    case 4: Foo<4>(j); break;
}

缺点是缓存压力，因此这只会在深度或长期运行的调用树中获得，其中值在持续时间内是恒定的。