250     f16x8_t        acc0, acc1, acc2, acc3;  in arm_mat_cmplx_mult_f16_4x4_mve()  local
277     acc3 = vcmulq(vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
278     acc3 = vcmlaq_rot90(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
284     mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]);  in arm_mat_cmplx_mult_f16_4x4_mve()
306     acc3 = vcmulq(vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
307     acc3 = vcmlaq_rot90(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
313     mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]);  in arm_mat_cmplx_mult_f16_4x4_mve()
335     acc3 = vcmulq(vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
336     acc3 = vcmlaq_rot90(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
342     mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]);  in arm_mat_cmplx_mult_f16_4x4_mve()
364     acc3 = vcmulq(vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
365     acc3 = vcmlaq_rot90(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16_4x4_mve()
371     mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]);  in arm_mat_cmplx_mult_f16_4x4_mve()
483             f16x8_t acc0, acc1, acc2, acc3;  in arm_mat_cmplx_mult_f16()  local
488             acc3 = vdupq_n_f16(0.0f16);  in arm_mat_cmplx_mult_f16()
524                 acc3 = vcmlaq(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16()
525                 acc3 = vcmlaq_rot90(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16()
561                 acc3 = vcmlaq(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16()
562                 acc3 = vcmlaq_rot90(acc3, vecA, vecB);  in arm_mat_cmplx_mult_f16()
570             mve_cmplx_sum_intra_vec_f16(acc3, &px[3 * CMPLX_DIM * numColsB + 0]);  in arm_mat_cmplx_mult_f16()