c++ - 汇编/__asm 内联

Question

我正在学习汇编并在我的 Digital Mars C++ 编译器中进行一些内联​​。我搜索了一些东西来使程序更好，并使用这些参数来调整程序:

use better C++ compiler//thinking of GCC or intel compiler

use assembly only in critical part of program 

find better algorithm

Cache miss, cache contention.

Loop-carried dependency chain.

Instruction fetching time.

Instruction decoding time.

Instruction retirement.

Register read stalls.

Execution port throughput.

Execution unit throughput.

Suboptimal reordering and scheduling of micro-ops.

Branch misprediction.

Floating point exception.

除了“register read stalls”，我都听懂了。

问题:谁能告诉我这在 CPU 中是如何发生的以及“乱序执行”的“超标量”形式？正常的“乱序”似乎合乎逻辑，但我找不到“超标量”形式的合乎逻辑的解释。

问题 2:有人还可以给出一些 SSE SSE2 和较新 CPU 的良好指令列表，最好是微操作表、端口吞吐量、单元和一些延迟计算表，以找到一段代码的真正瓶颈？

我会很高兴有这样一个小例子:

//loop carried dependency chain breaking:
__asm
{
loop_begin:
....
.... 
sub edx,05h //rather than taking i*5 in each iteration, we sub 5 each iteration
sub ecx,01h //i-- counter
...
...
jnz loop_begin//edit: sub ecx must have been after the sub edx for jnz
}
//while sub edx makes us get rid of a multiplication also makes that independent of ecx, making independent

谢谢。

计算机:Pentium-M 2GHz，Windows XP-32 位

Answer 1

你应该看看 Agner Fogs 优化手册:Optimizing software in C++: An optimization guide for Windows, Linux and Mac platforms或 Optimizing subroutines in assembly language: An optimization guide for x86 platforms .

但要真正胜过现代编译器，您需要对要优化的架构有一些良好的背景知识:The microarchitecture of Intel, AMD and VIA CPUs: An optimization guide for assembly programmers and compiler makers