1 | /* ----------------------------------------------------------------------
|
---|
2 | * Project: CMSIS DSP Library
|
---|
3 | * Title: arm_scale_q15.c
|
---|
4 | * Description: Multiplies a Q15 vector by a scalar
|
---|
5 | *
|
---|
6 | * $Date: 27. January 2017
|
---|
7 | * $Revision: V.1.5.1
|
---|
8 | *
|
---|
9 | * Target Processor: Cortex-M cores
|
---|
10 | * -------------------------------------------------------------------- */
|
---|
11 | /*
|
---|
12 | * Copyright (C) 2010-2017 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 "arm_math.h"
|
---|
30 |
|
---|
31 | /**
|
---|
32 | * @ingroup groupMath
|
---|
33 | */
|
---|
34 |
|
---|
35 | /**
|
---|
36 | * @addtogroup scale
|
---|
37 | * @{
|
---|
38 | */
|
---|
39 |
|
---|
40 | /**
|
---|
41 | * @brief Multiplies a Q15 vector by a scalar.
|
---|
42 | * @param[in] *pSrc points to the input vector
|
---|
43 | * @param[in] scaleFract fractional portion of the scale value
|
---|
44 | * @param[in] shift number of bits to shift the result by
|
---|
45 | * @param[out] *pDst points to the output vector
|
---|
46 | * @param[in] blockSize number of samples in the vector
|
---|
47 | * @return none.
|
---|
48 | *
|
---|
49 | * <b>Scaling and Overflow Behavior:</b>
|
---|
50 | * \par
|
---|
51 | * The input data <code>*pSrc</code> and <code>scaleFract</code> are in 1.15 format.
|
---|
52 | * These are multiplied to yield a 2.30 intermediate result and this is shifted with saturation to 1.15 format.
|
---|
53 | */
|
---|
54 |
|
---|
55 |
|
---|
56 | void arm_scale_q15(
|
---|
57 | q15_t * pSrc,
|
---|
58 | q15_t scaleFract,
|
---|
59 | int8_t shift,
|
---|
60 | q15_t * pDst,
|
---|
61 | uint32_t blockSize)
|
---|
62 | {
|
---|
63 | int8_t kShift = 15 - shift; /* shift to apply after scaling */
|
---|
64 | uint32_t blkCnt; /* loop counter */
|
---|
65 |
|
---|
66 | #if defined (ARM_MATH_DSP)
|
---|
67 |
|
---|
68 | /* Run the below code for Cortex-M4 and Cortex-M3 */
|
---|
69 | q15_t in1, in2, in3, in4;
|
---|
70 | q31_t inA1, inA2; /* Temporary variables */
|
---|
71 | q31_t out1, out2, out3, out4;
|
---|
72 |
|
---|
73 |
|
---|
74 | /*loop Unrolling */
|
---|
75 | blkCnt = blockSize >> 2U;
|
---|
76 |
|
---|
77 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
|
---|
78 | ** a second loop below computes the remaining 1 to 3 samples. */
|
---|
79 | while (blkCnt > 0U)
|
---|
80 | {
|
---|
81 | /* Reading 2 inputs from memory */
|
---|
82 | inA1 = *__SIMD32(pSrc)++;
|
---|
83 | inA2 = *__SIMD32(pSrc)++;
|
---|
84 |
|
---|
85 | /* C = A * scale */
|
---|
86 | /* Scale the inputs and then store the 2 results in the destination buffer
|
---|
87 | * in single cycle by packing the outputs */
|
---|
88 | out1 = (q31_t) ((q15_t) (inA1 >> 16) * scaleFract);
|
---|
89 | out2 = (q31_t) ((q15_t) inA1 * scaleFract);
|
---|
90 | out3 = (q31_t) ((q15_t) (inA2 >> 16) * scaleFract);
|
---|
91 | out4 = (q31_t) ((q15_t) inA2 * scaleFract);
|
---|
92 |
|
---|
93 | /* apply shifting */
|
---|
94 | out1 = out1 >> kShift;
|
---|
95 | out2 = out2 >> kShift;
|
---|
96 | out3 = out3 >> kShift;
|
---|
97 | out4 = out4 >> kShift;
|
---|
98 |
|
---|
99 | /* saturate the output */
|
---|
100 | in1 = (q15_t) (__SSAT(out1, 16));
|
---|
101 | in2 = (q15_t) (__SSAT(out2, 16));
|
---|
102 | in3 = (q15_t) (__SSAT(out3, 16));
|
---|
103 | in4 = (q15_t) (__SSAT(out4, 16));
|
---|
104 |
|
---|
105 | /* store the result to destination */
|
---|
106 | *__SIMD32(pDst)++ = __PKHBT(in2, in1, 16);
|
---|
107 | *__SIMD32(pDst)++ = __PKHBT(in4, in3, 16);
|
---|
108 |
|
---|
109 | /* Decrement the loop counter */
|
---|
110 | blkCnt--;
|
---|
111 | }
|
---|
112 |
|
---|
113 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
|
---|
114 | ** No loop unrolling is used. */
|
---|
115 | blkCnt = blockSize % 0x4U;
|
---|
116 |
|
---|
117 | while (blkCnt > 0U)
|
---|
118 | {
|
---|
119 | /* C = A * scale */
|
---|
120 | /* Scale the input and then store the result in the destination buffer. */
|
---|
121 | *pDst++ = (q15_t) (__SSAT(((*pSrc++) * scaleFract) >> kShift, 16));
|
---|
122 |
|
---|
123 | /* Decrement the loop counter */
|
---|
124 | blkCnt--;
|
---|
125 | }
|
---|
126 |
|
---|
127 | #else
|
---|
128 |
|
---|
129 | /* Run the below code for Cortex-M0 */
|
---|
130 |
|
---|
131 | /* Initialize blkCnt with number of samples */
|
---|
132 | blkCnt = blockSize;
|
---|
133 |
|
---|
134 | while (blkCnt > 0U)
|
---|
135 | {
|
---|
136 | /* C = A * scale */
|
---|
137 | /* Scale the input and then store the result in the destination buffer. */
|
---|
138 | *pDst++ = (q15_t) (__SSAT(((q31_t) * pSrc++ * scaleFract) >> kShift, 16));
|
---|
139 |
|
---|
140 | /* Decrement the loop counter */
|
---|
141 | blkCnt--;
|
---|
142 | }
|
---|
143 |
|
---|
144 | #endif /* #if defined (ARM_MATH_DSP) */
|
---|
145 |
|
---|
146 | }
|
---|
147 |
|
---|
148 | /**
|
---|
149 | * @} end of scale group
|
---|
150 | */
|
---|