1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_dot_prod_f64.c
4 * Description: Floating-point dot product
5 *
6 * $Date: 03 June 2022
7 * $Revision: V1.10.1
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/basic_math_functions.h"
30
31 /**
32 @ingroup groupMath
33 */
34
35 /**
36 @addtogroup BasicDotProd
37 @{
38 */
39
40 /**
41 @brief Dot product of floating-point vectors.
42 @param[in] pSrcA points to the first input vector.
43 @param[in] pSrcB points to the second input vector.
44 @param[in] blockSize number of samples in each vector.
45 @param[out] result output result returned here.
46 */
47 #if defined(ARM_MATH_NEON) && defined(__aarch64__)
arm_dot_prod_f64(const float64_t * pSrcA,const float64_t * pSrcB,uint32_t blockSize,float64_t * result)48 ARM_DSP_ATTRIBUTE void arm_dot_prod_f64(
49 const float64_t * pSrcA,
50 const float64_t * pSrcB,
51 uint32_t blockSize,
52 float64_t * result)
53 {
54 uint32_t blkCnt; /* Loop counter */
55 float64_t sum = 0.; /* Temporary return variable */
56 /* Neon Buffer Initialisation */
57 float64x2_t sumV = vdupq_n_f64(0.0); /* Neon buffer for sum variable */
58
59
60 /* Neon Buffer for sources */
61 float64x2_t pSrcAV;
62 float64x2_t pSrcBV;
63
64 /* Initialize blkCnt with number of samples */
65 blkCnt = blockSize >> 1U;
66
67 while (blkCnt > 0U)
68 {
69 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
70
71 /* Load source value in Neon Buffer */
72 pSrcAV = vld1q_f64(pSrcA);
73 pSrcBV = vld1q_f64(pSrcB);
74 /* Calculate dot product and store result in a temporary buffer. */
75 sumV = vmlaq_f64(sumV, pSrcAV, pSrcBV);
76
77 pSrcA+=2;
78 pSrcB+=2;
79 /* Decrement loop counter */
80 blkCnt--;
81 }
82 /* Sum both 64 bits part in the float64x2 */
83 sum = vaddvq_f64(sumV);
84
85
86 /* Tail */
87 blkCnt = blockSize & 1 ;
88
89 while(blkCnt > 0U)
90 {
91 sum += (*pSrcA++) * (*pSrcB++);
92
93 /* Decrement loop counter */
94 blkCnt--;
95 }
96
97 /* Store result in destination buffer */
98 *result = sum;
99 }
100 #else
arm_dot_prod_f64(const float64_t * pSrcA,const float64_t * pSrcB,uint32_t blockSize,float64_t * result)101 ARM_DSP_ATTRIBUTE void arm_dot_prod_f64(
102 const float64_t * pSrcA,
103 const float64_t * pSrcB,
104 uint32_t blockSize,
105 float64_t * result)
106 {
107 uint32_t blkCnt; /* Loop counter */
108 float64_t sum = 0.; /* Temporary return variable */
109
110 /* Initialize blkCnt with number of samples */
111 blkCnt = blockSize;
112
113 while (blkCnt > 0U)
114 {
115 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
116
117 /* Calculate dot product and store result in a temporary buffer. */
118 sum += (*pSrcA++) * (*pSrcB++);
119
120 /* Decrement loop counter */
121 blkCnt--;
122 }
123
124 /* Store result in destination buffer */
125 *result = sum;
126 }
127 #endif
128
129 /**
130 @} end of BasicDotProd group
131 */
132
