Rev 2 | Details | Compare with Previous | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
2 | mjames | 1 | /* ---------------------------------------------------------------------- |
2 | * Project: CMSIS DSP Library |
||
3 | * Title: arm_mat_sub_f32.c |
||
4 | * Description: Floating-point matrix subtraction |
||
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 groupMatrix |
||
33 | */ |
||
34 | |||
35 | /** |
||
36 | * @defgroup MatrixSub Matrix Subtraction |
||
37 | * |
||
38 | * Subtract two matrices. |
||
39 | * \image html MatrixSubtraction.gif "Subraction of two 3 x 3 matrices" |
||
40 | * |
||
41 | * The functions check to make sure that |
||
42 | * <code>pSrcA</code>, <code>pSrcB</code>, and <code>pDst</code> have the same |
||
43 | * number of rows and columns. |
||
44 | */ |
||
45 | |||
46 | /** |
||
47 | * @addtogroup MatrixSub |
||
48 | * @{ |
||
49 | */ |
||
50 | |||
51 | /** |
||
52 | * @brief Floating-point matrix subtraction |
||
53 | * @param[in] *pSrcA points to the first input matrix structure |
||
54 | * @param[in] *pSrcB points to the second input matrix structure |
||
55 | * @param[out] *pDst points to output matrix structure |
||
56 | * @return The function returns either |
||
57 | * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking. |
||
58 | */ |
||
59 | |||
60 | arm_status arm_mat_sub_f32( |
||
61 | const arm_matrix_instance_f32 * pSrcA, |
||
62 | const arm_matrix_instance_f32 * pSrcB, |
||
63 | arm_matrix_instance_f32 * pDst) |
||
64 | { |
||
65 | float32_t *pIn1 = pSrcA->pData; /* input data matrix pointer A */ |
||
66 | float32_t *pIn2 = pSrcB->pData; /* input data matrix pointer B */ |
||
67 | float32_t *pOut = pDst->pData; /* output data matrix pointer */ |
||
68 | |||
69 | #if defined (ARM_MATH_DSP) |
||
70 | |||
71 | float32_t inA1, inA2, inB1, inB2, out1, out2; /* temporary variables */ |
||
72 | |||
73 | #endif // #if defined (ARM_MATH_DSP) |
||
74 | |||
75 | uint32_t numSamples; /* total number of elements in the matrix */ |
||
76 | uint32_t blkCnt; /* loop counters */ |
||
77 | arm_status status; /* status of matrix subtraction */ |
||
78 | |||
79 | #ifdef ARM_MATH_MATRIX_CHECK |
||
80 | /* Check for matrix mismatch condition */ |
||
81 | if ((pSrcA->numRows != pSrcB->numRows) || |
||
82 | (pSrcA->numCols != pSrcB->numCols) || |
||
83 | (pSrcA->numRows != pDst->numRows) || (pSrcA->numCols != pDst->numCols)) |
||
84 | { |
||
85 | /* Set status as ARM_MATH_SIZE_MISMATCH */ |
||
86 | status = ARM_MATH_SIZE_MISMATCH; |
||
87 | } |
||
88 | else |
||
89 | #endif /* #ifdef ARM_MATH_MATRIX_CHECK */ |
||
90 | { |
||
91 | /* Total number of samples in the input matrix */ |
||
92 | numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; |
||
93 | |||
94 | #if defined (ARM_MATH_DSP) |
||
95 | |||
96 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
97 | |||
98 | /* Loop Unrolling */ |
||
99 | blkCnt = numSamples >> 2U; |
||
100 | |||
101 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
102 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
103 | while (blkCnt > 0U) |
||
104 | { |
||
105 | /* C(m,n) = A(m,n) - B(m,n) */ |
||
106 | /* Subtract and then store the results in the destination buffer. */ |
||
107 | /* Read values from source A */ |
||
108 | inA1 = pIn1[0]; |
||
109 | |||
110 | /* Read values from source B */ |
||
111 | inB1 = pIn2[0]; |
||
112 | |||
113 | /* Read values from source A */ |
||
114 | inA2 = pIn1[1]; |
||
115 | |||
116 | /* out = sourceA - sourceB */ |
||
117 | out1 = inA1 - inB1; |
||
118 | |||
119 | /* Read values from source B */ |
||
120 | inB2 = pIn2[1]; |
||
121 | |||
122 | /* Read values from source A */ |
||
123 | inA1 = pIn1[2]; |
||
124 | |||
125 | /* out = sourceA - sourceB */ |
||
126 | out2 = inA2 - inB2; |
||
127 | |||
128 | /* Read values from source B */ |
||
129 | inB1 = pIn2[2]; |
||
130 | |||
131 | /* Store result in destination */ |
||
132 | pOut[0] = out1; |
||
133 | pOut[1] = out2; |
||
134 | |||
135 | /* Read values from source A */ |
||
136 | inA2 = pIn1[3]; |
||
137 | |||
138 | /* Read values from source B */ |
||
139 | inB2 = pIn2[3]; |
||
140 | |||
141 | /* out = sourceA - sourceB */ |
||
142 | out1 = inA1 - inB1; |
||
143 | |||
144 | |||
145 | /* out = sourceA - sourceB */ |
||
146 | out2 = inA2 - inB2; |
||
147 | |||
148 | /* Store result in destination */ |
||
149 | pOut[2] = out1; |
||
150 | |||
151 | /* Store result in destination */ |
||
152 | pOut[3] = out2; |
||
153 | |||
154 | |||
155 | /* update pointers to process next sampels */ |
||
156 | pIn1 += 4U; |
||
157 | pIn2 += 4U; |
||
158 | pOut += 4U; |
||
159 | |||
160 | /* Decrement the loop counter */ |
||
161 | blkCnt--; |
||
162 | } |
||
163 | |||
164 | /* If the numSamples is not a multiple of 4, compute any remaining output samples here. |
||
165 | ** No loop unrolling is used. */ |
||
166 | blkCnt = numSamples % 0x4U; |
||
167 | |||
168 | #else |
||
169 | |||
170 | /* Run the below code for Cortex-M0 */ |
||
171 | |||
172 | /* Initialize blkCnt with number of samples */ |
||
173 | blkCnt = numSamples; |
||
174 | |||
175 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
176 | |||
177 | while (blkCnt > 0U) |
||
178 | { |
||
179 | /* C(m,n) = A(m,n) - B(m,n) */ |
||
180 | /* Subtract and then store the results in the destination buffer. */ |
||
181 | *pOut++ = (*pIn1++) - (*pIn2++); |
||
182 | |||
183 | /* Decrement the loop counter */ |
||
184 | blkCnt--; |
||
185 | } |
||
186 | |||
187 | /* Set status as ARM_MATH_SUCCESS */ |
||
188 | status = ARM_MATH_SUCCESS; |
||
189 | } |
||
190 | |||
191 | /* Return to application */ |
||
192 | return (status); |
||
193 | } |
||
194 | |||
195 | /** |
||
196 | * @} end of MatrixSub group |
||
197 | */ |