1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_fill_f16.c
4 * Description: Fills a constant value into 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 Fill
41 @{
42 */
43
44 /**
45 @brief Fills a constant value into a f16 vector.
46 @param[in] value input value to be filled
47 @param[out] pDst points to output vector
48 @param[in] blockSize number of samples in each vector
49 */
50 #if defined(ARM_MATH_MVE_FLOAT16) && !defined(ARM_MATH_AUTOVECTORIZE)
51
arm_fill_f16(float16_t value,float16_t * pDst,uint32_t blockSize)52 ARM_DSP_ATTRIBUTE void arm_fill_f16(
53 float16_t value,
54 float16_t * pDst,
55 uint32_t blockSize)
56 {
57 do {
58 mve_pred16_t p = vctp16q(blockSize);
59
60 vstrhq_p_f16(pDst,
61 vdupq_m_n_f16(vuninitializedq_f16(), value, p), p);
62 /*
63 * Decrement the blockSize loop counter
64 * Advance vector source and destination pointers
65 */
66 pDst += 8;
67 blockSize -= 8;
68 }
69 while ((int32_t) blockSize > 0);
70 }
71 #else
arm_fill_f16(float16_t value,float16_t * pDst,uint32_t blockSize)72 ARM_DSP_ATTRIBUTE void arm_fill_f16(
73 float16_t value,
74 float16_t * pDst,
75 uint32_t blockSize)
76 {
77 uint32_t blkCnt; /* Loop counter */
78
79 #if defined (ARM_MATH_LOOPUNROLL)
80
81 /* Loop unrolling: Compute 4 outputs at a time */
82 blkCnt = blockSize >> 2U;
83
84 while (blkCnt > 0U)
85 {
86 /* C = value */
87
88 /* Fill value in destination buffer */
89 *pDst++ = value;
90 *pDst++ = value;
91 *pDst++ = value;
92 *pDst++ = value;
93
94 /* Decrement loop counter */
95 blkCnt--;
96 }
97
98 /* Loop unrolling: Compute remaining outputs */
99 blkCnt = blockSize % 0x4U;
100
101 #else
102
103 /* Initialize blkCnt with number of samples */
104 blkCnt = blockSize;
105
106 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
107
108 while (blkCnt > 0U)
109 {
110 /* C = value */
111
112 /* Fill value in destination buffer */
113 *pDst++ = value;
114
115 /* Decrement loop counter */
116 blkCnt--;
117 }
118 }
119 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
120
121 /**
122 @} end of Fill group
123 */
124
125 #endif /* #if defined(ARM_FLOAT16_SUPPORTED) */
126
127
