void test_vsubQs8 (void)
{
int8x16_t out_int8x16_t;
int8x16_t arg0_int8x16_t;
int8x16_t arg1_int8x16_t;
out_int8x16_t = vsubq_s8 (arg0_int8x16_t, arg1_int8x16_t);
}
/* s8x16 sub */
void mw_neon_mm_sub_s8x16(signed char * A, int Row, int Col, signed char * B, signed char * C)
{
int8x16_t neon_a, neon_b, neon_c;
int size = Row * Col;
int i = 0;
int k = 0;
for (i = 16; i <= size ; i+=16)
{
k = i - 16;
neon_a = vld1q_s8(A + k);
neon_b = vld1q_s8(B + k);
neon_c = vsubq_s8(neon_a, neon_b);
vst1q_s8(C + k, neon_c);
}
k = i - 16;
for (i = 0; i <= size % 16; i++)
{
C[k + i] = A[k + i] - B[k + i];
}
}
inline int8x16_t vsubq(const int8x16_t & v0, const int8x16_t & v1) { return vsubq_s8 (v0, v1); }
