void test_vqaddQu16 (void)
{
uint16x8_t out_uint16x8_t;
uint16x8_t arg0_uint16x8_t;
uint16x8_t arg1_uint16x8_t;
out_uint16x8_t = vqaddq_u16 (arg0_uint16x8_t, arg1_uint16x8_t);
}
/* u16x8 saturated add */
void mw_neon_mm_qadd_u16x8(unsigned short * A, int Row, int Col, unsigned short * B, unsigned short * C)
{
uint16x8_t neon_a, neon_b, neon_c;
int size = Row * Col;
int i = 0;
int k = 0;
for (i = 8; i <= size ; i+=8)
{
k = i - 8;
neon_a = vld1q_u16(A + k);
neon_b = vld1q_u16(B + k);
neon_c = vqaddq_u16(neon_a, neon_b);
vst1q_u16(C + k, neon_c);
}
k = i - 8;
for (i = 0; i < size % 8; i++)
{
C[k + i] = A[k + i] + B[k + i];
}
}
inline uint16x8_t vqaddq(const uint16x8_t & v0, const uint16x8_t & v1) { return vqaddq_u16(v0, v1); }
