Effective SIMD Vectorization for Intel Xeon Phi Coprocessors

Table 3

_v512 after zero initialization.

A_v512	0	0	0	0
	0	0	0	0
	0	0	0	0
	0	0	0	0