1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_copy_f16.c
4 * Description: Copies the elements of a floating-point vector
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/support_functions_f16.h"
30
31 #if defined(ARM_FLOAT16_SUPPORTED)
32
33
34 /**
35 @ingroup groupSupport
36 */
37
38
39 /**
40 @addtogroup copy
41 @{
42 */
43
44 /**
45 @brief Copies the elements of a f16 vector.
46 @param[in] pSrc points to input vector
47 @param[out] pDst points to output vector
48 @param[in] blockSize number of samples in each vector
49 @return none
50 */
51 #if defined(ARM_MATH_MVE_FLOAT16) && !defined(ARM_MATH_AUTOVECTORIZE)
52
arm_copy_f16(const float16_t * pSrc,float16_t * pDst,uint32_t blockSize)53 void arm_copy_f16(
54 const float16_t * pSrc,
55 float16_t * pDst,
56 uint32_t blockSize)
57 {
58 do {
59 mve_pred16_t p = vctp16q(blockSize);
60
61 vstrhq_p_f16(pDst,
62 vldrhq_z_f16((float16_t const *) pSrc, p), p);
63 /*
64 * Decrement the blockSize loop counter
65 * Advance vector source and destination pointers
66 */
67 pSrc += 8;
68 pDst += 8;
69 blockSize -= 8;
70 }
71 while ((int32_t) blockSize > 0);
72 }
73
74 #else
75
arm_copy_f16(const float16_t * pSrc,float16_t * pDst,uint32_t blockSize)76 void arm_copy_f16(
77 const float16_t * pSrc,
78 float16_t * pDst,
79 uint32_t blockSize)
80 {
81 uint32_t blkCnt; /* Loop counter */
82
83 #if defined (ARM_MATH_LOOPUNROLL)
84
85 /* Loop unrolling: Compute 4 outputs at a time */
86 blkCnt = blockSize >> 2U;
87
88 while (blkCnt > 0U)
89 {
90 /* C = A */
91
92 /* Copy and store result in destination buffer */
93 *pDst++ = *pSrc++;
94 *pDst++ = *pSrc++;
95 *pDst++ = *pSrc++;
96 *pDst++ = *pSrc++;
97
98 /* Decrement loop counter */
99 blkCnt--;
100 }
101
102 /* Loop unrolling: Compute remaining outputs */
103 blkCnt = blockSize % 0x4U;
104
105 #else
106
107 /* Initialize blkCnt with number of samples */
108 blkCnt = blockSize;
109
110 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
111
112 while (blkCnt > 0U)
113 {
114 /* C = A */
115
116 /* Copy and store result in destination buffer */
117 *pDst++ = *pSrc++;
118
119 /* Decrement loop counter */
120 blkCnt--;
121 }
122 }
123 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
124
125 /**
126 @} end of BasicCopy group
127 */
128
129 #endif /* #if defined(ARM_FLOAT16_SUPPORTED) */
130
131
