First of all, before we submerge into the world of macro instructions, we have to understand what they actually are. Putting it the simplest way, macro instructions are aliases for sequences of instructions. You may be familiar with the term from high-level languages ( we say "may be" because not all high-level languages implement this feature), but we'll still explain it here. Remember the following sequence from the previous chapter?

movd xmm3, [dpy]
movlhps xmm3, xmm3
movsldup xmm3, xmm3

This sequence loads all four singles of an XMM register (in this specific case, it was XMM3) with a single precision floating point value from memory pointed by dpy. We used such sequences several times in our code, so it would be natural to try and replace it with a single macro instruction. Thus, defining the following macro would make our code look more elegant and readable:

macro load_4 xmmreg, addr
{
   movd xmmreg, [addr]
   movlhps xmmreg, xmmreg
   movsldup xmmreg, xmmreg
}

We use it in our code like this:

load_4 xmm3, dpy
load_4 xmm4, pi_2

This would make the code look more elegant and much more readable.