This will not necessarily help you write efficient code!
i86 op codes are more or less a "legacy" format that persists because of the sheer volume of code and executable binaries for Windows and Linux out there.
Its a bit like the old scholars writing in latin, an Italian speaker like Galileo would write in Latin and his paper could be understood by a Polish speaker like Copernicus. This was still the most effective way to communicate even though niether was particulary good at Latin, and Latin is a rubbish language for expressing mathematical ideas.
So compilers generate x86 code by default, and, modern chips read the anceint Op codes and transalate what they see into parallel risc instructions, with reordered execution, speculative execution, pipelining etc. plus they make full use of the 32 or 64 registers the processor actually has (as opposed to the pathetic 8 you see in x86 instructions.)
Now all optimising compilers know this is what really happens, so they code up sequences of OP codes which they know the chip can optimise efficiently -- even though some of these sequences would look innefficient to an circa 1990 .asm programmer.
At some point you need to accept that the 10s of thousands of man years effort compiler writers have put in have paid off, and, trust them.
The simplest and easiest way to get a more eficient runtime is to buy the Intel C/C++ compiler. They have a niche market for efficeint compilers, and, they have the advantage of being able to ask the chip designers about what goes on inside.
