movups x(,%eax,4), %xmm1 mulps %xmm1, %xmm2 Intel® AVX enables these two instructions to be combined. vmulps x(,%eax,4), %xmm2, %xmm2 In Part 2: Effectively using assembly language listings, and code ...