Effective SIMD Vectorization for Intel Xeon Phi Coprocessors

Table 4

Vector register contents after first shuffle.

A_v512	A11	0	0	0
	0	A22	0	0
	0	0	A33	0
	0	0	0	A44