1  /* ----------------------------------------------------------------------
  2   * Project:      CMSIS DSP Library
  3   * Title:        arm_quaternion_product_f32.c
  4   * Description:  Floating-point quaternion product
  5   *
  6   * $Date:        23 April 2021
  7   * $Revision:    V1.9.0
  8   *
  9   * Target Processor: Cortex-M and Cortex-A cores
 10   * -------------------------------------------------------------------- */
 11  /*
 12   * Copyright (C) 2010-2021 ARM Limited or its affiliates. All rights reserved.
 13   *
 14   * SPDX-License-Identifier: Apache-2.0
 15   *
 16   * Licensed under the Apache License, Version 2.0 (the License); you may
 17   * not use this file except in compliance with the License.
 18   * You may obtain a copy of the License at
 19   *
 20   * www.apache.org/licenses/LICENSE-2.0
 21   *
 22   * Unless required by applicable law or agreed to in writing, software
 23   * distributed under the License is distributed on an AS IS BASIS, WITHOUT
 24   * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 25   * See the License for the specific language governing permissions and
 26   * limitations under the License.
 27   */
 28  
 29  #include "dsp/quaternion_math_functions.h"
 30  #include <math.h>
 31  
 32  /**
 33    @ingroup groupQuaternionMath
 34   */
 35  
 36  /**
 37    @defgroup QuatProd Quaternion Product
 38  
 39    Compute the product of quaternions.
 40   */
 41  
 42  /**
 43    @ingroup QuatProd
 44   */
 45  
 46  /**
 47    @defgroup QuatProdVect Elementwise Quaternion Product
 48  
 49    Compute the elementwise product of quaternions.
 50   */
 51  
 52  /**
 53    @addtogroup QuatProdVect
 54    @{
 55   */
 56  
 57  /**
 58    @brief         Floating-point elementwise product two quaternions.
 59    @param[in]     qa                  first array of quaternions
 60    @param[in]     qb                  second array of quaternions
 61    @param[out]    qr                   elementwise product of quaternions
 62    @param[in]     nbQuaternions       number of quaternions in the array
 63    @return        none
 64   */
 65  
 66  #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
 67  
 68  #include "arm_helium_utils.h"
 69  
 70  void arm_quaternion_product_f32(const float32_t *qa, 
 71      const float32_t *qb, 
 72      float32_t *qr,
 73      uint32_t nbQuaternions)
 74  {
 75      static uint32_t patternA[4] = { 0, 1, 0, 1 };
 76      static uint32_t patternB[4] = { 3, 2, 3, 2 };
 77      static uint32_t patternC[4] = { 3, 2, 1, 0 };
 78      static float32_t   signA[4] = { -1, -1, 1, 1 };
 79  
 80      uint32x4_t vecA = vld1q_u32(patternA);
 81      uint32x4_t vecB = vld1q_u32(patternB);
 82      uint32x4_t vecC = vld1q_u32(patternC);
 83      f32x4_t vecSignA = vld1q_f32(signA);
 84  
 85      while (nbQuaternions > 0U)
 86      {
 87          f32x4_t vecTmpA, vecTmpB, vecAcc;
 88  
 89          vecTmpA = vldrwq_gather_shifted_offset_f32(qa, vecA);
 90          vecTmpB = vld1q(qb);
 91          /*
 92           * vcmul(r, [a1, a2, a1, a2], [b1, b2, b3, b4], 0)
 93           */
 94          vecAcc = vcmulq(vecTmpA, vecTmpB);
 95          /*
 96           * vcmla(r, [a1, a2, a1, a2], [b1, b2, b3, b4], 90)
 97           */
 98          vecAcc = vcmlaq_rot90(vecAcc, vecTmpA, vecTmpB);
 99  
100          vecTmpA = vldrwq_gather_shifted_offset_f32(qa, vecB);
101          vecTmpB = vldrwq_gather_shifted_offset_f32(qb, vecC);
102          /*
103           * build [-b4, -b3, b2, b1]
104           */
105          vecTmpB = vecTmpB * vecSignA;
106          /*
107           * vcmla(r, [a4, a3, a4, a3], [-b4, -b3, b2, b1], 270)
108           */
109          vecAcc = vcmlaq_rot270(vecAcc, vecTmpA, vecTmpB);
110          /*
111           * vcmla(r, [a4, a3, a4, a3], [-b4, -b3, b2, b1], 0)
112           */
113          vecAcc = vcmlaq(vecAcc, vecTmpA, vecTmpB);
114          /*
115           * store accumulator
116           */
117          vst1q_f32(qr, vecAcc);
118  
119          /* move to next quaternion */
120          qa += 4;
121          qb += 4;
122          qr += 4;
123  
124          nbQuaternions--;
125      }
126  }
127  
128  #else
129  
130  void arm_quaternion_product_f32(const float32_t *qa, 
131      const float32_t *qb, 
132      float32_t *qr,
133      uint32_t nbQuaternions)
134  {
135     uint32_t i;
136     for(i=0; i < nbQuaternions; i++)
137     {
138       arm_quaternion_product_single_f32(qa, qb, qr);
139  
140       qa += 4;
141       qb += 4;
142       qr += 4;
143     }
144  }
145  #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
146  
147  /**
148    @} end of QuatProdVect group
149   */