1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_negate_f32.c
4 * Description: Negates floating-point vectors
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/basic_math_functions.h"
30
31 /**
32 @ingroup groupMath
33 */
34
35 /**
36 @defgroup BasicNegate Vector Negate
37
38 Negates the elements of a vector.
39
40 <pre>
41 pDst[n] = -pSrc[n], 0 <= n < blockSize.
42 </pre>
43
44 The functions support in-place computation allowing the source and
45 destination pointers to reference the same memory buffer.
46 There are separate functions for floating-point, Q7, Q15, and Q31 data types.
47 */
48
49 /**
50 @addtogroup BasicNegate
51 @{
52 */
53
54 /**
55 @brief Negates the elements of a floating-point vector.
56 @param[in] pSrc points to input vector.
57 @param[out] pDst points to output vector.
58 @param[in] blockSize number of samples in each vector.
59 */
60
61 #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
62
63 #include "arm_helium_utils.h"
64
arm_negate_f32(const float32_t * pSrc,float32_t * pDst,uint32_t blockSize)65 void arm_negate_f32(
66 const float32_t * pSrc,
67 float32_t * pDst,
68 uint32_t blockSize)
69 {
70 uint32_t blkCnt; /* Loop counter */
71 f32x4_t vec1;
72 f32x4_t res;
73
74
75 /* Compute 4 outputs at a time */
76 blkCnt = blockSize >> 2U;
77 while (blkCnt > 0U)
78 {
79 /* C = |A| */
80
81 /* Calculate absolute values and then store the results in the destination buffer. */
82 vec1 = vld1q(pSrc);
83 res = vnegq(vec1);
84 vst1q(pDst, res);
85
86 /* Increment pointers */
87 pSrc += 4;
88 pDst += 4;
89
90 /* Decrement the loop counter */
91 blkCnt--;
92 }
93
94 /* Tail */
95 blkCnt = blockSize & 0x3;
96 if (blkCnt > 0U)
97 {
98 /* C = |A| */
99 mve_pred16_t p0 = vctp32q(blkCnt);
100 vec1 = vld1q((float32_t const *) pSrc);
101 vstrwq_p(pDst, vnegq(vec1), p0);
102 }
103
104 }
105
106 #else
arm_negate_f32(const float32_t * pSrc,float32_t * pDst,uint32_t blockSize)107 void arm_negate_f32(
108 const float32_t * pSrc,
109 float32_t * pDst,
110 uint32_t blockSize)
111 {
112 uint32_t blkCnt; /* Loop counter */
113
114 #if defined(ARM_MATH_NEON_EXPERIMENTAL) && !defined(ARM_MATH_AUTOVECTORIZE)
115 f32x4_t vec1;
116 f32x4_t res;
117
118 /* Compute 4 outputs at a time */
119 blkCnt = blockSize >> 2U;
120
121 while (blkCnt > 0U)
122 {
123 /* C = -A */
124
125 /* Negate and then store the results in the destination buffer. */
126 vec1 = vld1q_f32(pSrc);
127 res = vnegq_f32(vec1);
128 vst1q_f32(pDst, res);
129
130 /* Increment pointers */
131 pSrc += 4;
132 pDst += 4;
133
134 /* Decrement the loop counter */
135 blkCnt--;
136 }
137
138 /* Tail */
139 blkCnt = blockSize & 0x3;
140
141 #else
142 #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
143
144 /* Loop unrolling: Compute 4 outputs at a time */
145 blkCnt = blockSize >> 2U;
146
147 while (blkCnt > 0U)
148 {
149 /* C = -A */
150
151 /* Negate and store result in destination buffer. */
152 *pDst++ = -*pSrc++;
153
154 *pDst++ = -*pSrc++;
155
156 *pDst++ = -*pSrc++;
157
158 *pDst++ = -*pSrc++;
159
160 /* Decrement loop counter */
161 blkCnt--;
162 }
163
164 /* Loop unrolling: Compute remaining outputs */
165 blkCnt = blockSize % 0x4U;
166
167 #else
168
169 /* Initialize blkCnt with number of samples */
170 blkCnt = blockSize;
171
172 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
173 #endif /* #if defined(ARM_MATH_NEON_EXPERIMENTAL) */
174
175 while (blkCnt > 0U)
176 {
177 /* C = -A */
178
179 /* Negate and store result in destination buffer. */
180 *pDst++ = -*pSrc++;
181
182 /* Decrement loop counter */
183 blkCnt--;
184 }
185
186 }
187 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
188
189 /**
190 @} end of BasicNegate group
191 */
192
