Rev 2 | Details | Compare with Previous | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 2 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Copyright (C) 2010-2014 ARM Limited. All rights reserved. |
||
| 3 | * |
||
| 4 | * $Date: 19. March 2015 |
||
| 5 | * $Revision: V.1.4.5 |
||
| 6 | * |
||
| 7 | * Project: CMSIS DSP Library |
||
| 8 | * Title: arm_mat_scale_q15.c |
||
| 9 | * |
||
| 10 | * Description: Multiplies a Q15 matrix by a scalar. |
||
| 11 | * |
||
| 12 | * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0 |
||
| 13 | * |
||
| 14 | * Redistribution and use in source and binary forms, with or without |
||
| 15 | * modification, are permitted provided that the following conditions |
||
| 16 | * are met: |
||
| 17 | * - Redistributions of source code must retain the above copyright |
||
| 18 | * notice, this list of conditions and the following disclaimer. |
||
| 19 | * - Redistributions in binary form must reproduce the above copyright |
||
| 20 | * notice, this list of conditions and the following disclaimer in |
||
| 21 | * the documentation and/or other materials provided with the |
||
| 22 | * distribution. |
||
| 23 | * - Neither the name of ARM LIMITED nor the names of its contributors |
||
| 24 | * may be used to endorse or promote products derived from this |
||
| 25 | * software without specific prior written permission. |
||
| 26 | * |
||
| 27 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
||
| 28 | * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
||
| 29 | * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS |
||
| 30 | * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE |
||
| 31 | * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
||
| 32 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, |
||
| 33 | * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; |
||
| 34 | * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
||
| 35 | * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
||
| 36 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN |
||
| 37 | * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
||
| 38 | * POSSIBILITY OF SUCH DAMAGE. |
||
| 39 | * -------------------------------------------------------------------- */ |
||
| 40 | |||
| 41 | #include "arm_math.h" |
||
| 42 | |||
| 43 | /** |
||
| 44 | * @ingroup groupMatrix |
||
| 45 | */ |
||
| 46 | |||
| 47 | /** |
||
| 48 | * @addtogroup MatrixScale |
||
| 49 | * @{ |
||
| 50 | */ |
||
| 51 | |||
| 52 | /** |
||
| 53 | * @brief Q15 matrix scaling. |
||
| 54 | * @param[in] *pSrc points to input matrix |
||
| 55 | * @param[in] scaleFract fractional portion of the scale factor |
||
| 56 | * @param[in] shift number of bits to shift the result by |
||
| 57 | * @param[out] *pDst points to output matrix structure |
||
| 58 | * @return The function returns either |
||
| 59 | * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking. |
||
| 60 | * |
||
| 61 | * @details |
||
| 62 | * <b>Scaling and Overflow Behavior:</b> |
||
| 63 | * \par |
||
| 64 | * The input data <code>*pSrc</code> and <code>scaleFract</code> are in 1.15 format. |
||
| 65 | * These are multiplied to yield a 2.30 intermediate result and this is shifted with saturation to 1.15 format. |
||
| 66 | */ |
||
| 67 | |||
| 68 | arm_status arm_mat_scale_q15( |
||
| 69 | const arm_matrix_instance_q15 * pSrc, |
||
| 70 | q15_t scaleFract, |
||
| 71 | int32_t shift, |
||
| 72 | arm_matrix_instance_q15 * pDst) |
||
| 73 | { |
||
| 74 | q15_t *pIn = pSrc->pData; /* input data matrix pointer */ |
||
| 75 | q15_t *pOut = pDst->pData; /* output data matrix pointer */ |
||
| 76 | uint32_t numSamples; /* total number of elements in the matrix */ |
||
| 77 | int32_t totShift = 15 - shift; /* total shift to apply after scaling */ |
||
| 78 | uint32_t blkCnt; /* loop counters */ |
||
| 79 | arm_status status; /* status of matrix scaling */ |
||
| 80 | |||
| 81 | #ifndef ARM_MATH_CM0_FAMILY |
||
| 82 | |||
| 83 | q15_t in1, in2, in3, in4; |
||
| 84 | q31_t out1, out2, out3, out4; |
||
| 85 | q31_t inA1, inA2; |
||
| 86 | |||
| 87 | #endif // #ifndef ARM_MATH_CM0_FAMILY |
||
| 88 | |||
| 89 | #ifdef ARM_MATH_MATRIX_CHECK |
||
| 90 | /* Check for matrix mismatch */ |
||
| 91 | if((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols)) |
||
| 92 | { |
||
| 93 | /* Set status as ARM_MATH_SIZE_MISMATCH */ |
||
| 94 | status = ARM_MATH_SIZE_MISMATCH; |
||
| 95 | } |
||
| 96 | else |
||
| 97 | #endif // #ifdef ARM_MATH_MATRIX_CHECK |
||
| 98 | { |
||
| 99 | /* Total number of samples in the input matrix */ |
||
| 100 | numSamples = (uint32_t) pSrc->numRows * pSrc->numCols; |
||
| 101 | |||
| 102 | #ifndef ARM_MATH_CM0_FAMILY |
||
| 103 | |||
| 104 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 105 | /* Loop Unrolling */ |
||
| 106 | blkCnt = numSamples >> 2; |
||
| 107 | |||
| 108 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 109 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 110 | while(blkCnt > 0u) |
||
| 111 | { |
||
| 112 | /* C(m,n) = A(m,n) * k */ |
||
| 113 | /* Scale, saturate and then store the results in the destination buffer. */ |
||
| 114 | /* Reading 2 inputs from memory */ |
||
| 115 | inA1 = _SIMD32_OFFSET(pIn); |
||
| 116 | inA2 = _SIMD32_OFFSET(pIn + 2); |
||
| 117 | |||
| 118 | /* C = A * scale */ |
||
| 119 | /* Scale the inputs and then store the 2 results in the destination buffer |
||
| 120 | * in single cycle by packing the outputs */ |
||
| 121 | out1 = (q31_t) ((q15_t) (inA1 >> 16) * scaleFract); |
||
| 122 | out2 = (q31_t) ((q15_t) inA1 * scaleFract); |
||
| 123 | out3 = (q31_t) ((q15_t) (inA2 >> 16) * scaleFract); |
||
| 124 | out4 = (q31_t) ((q15_t) inA2 * scaleFract); |
||
| 125 | |||
| 126 | out1 = out1 >> totShift; |
||
| 127 | inA1 = _SIMD32_OFFSET(pIn + 4); |
||
| 128 | out2 = out2 >> totShift; |
||
| 129 | inA2 = _SIMD32_OFFSET(pIn + 6); |
||
| 130 | out3 = out3 >> totShift; |
||
| 131 | out4 = out4 >> totShift; |
||
| 132 | |||
| 133 | in1 = (q15_t) (__SSAT(out1, 16)); |
||
| 134 | in2 = (q15_t) (__SSAT(out2, 16)); |
||
| 135 | in3 = (q15_t) (__SSAT(out3, 16)); |
||
| 136 | in4 = (q15_t) (__SSAT(out4, 16)); |
||
| 137 | |||
| 138 | _SIMD32_OFFSET(pOut) = __PKHBT(in2, in1, 16); |
||
| 139 | _SIMD32_OFFSET(pOut + 2) = __PKHBT(in4, in3, 16); |
||
| 140 | |||
| 141 | /* update pointers to process next sampels */ |
||
| 142 | pIn += 4u; |
||
| 143 | pOut += 4u; |
||
| 144 | |||
| 145 | |||
| 146 | /* Decrement the numSamples loop counter */ |
||
| 147 | blkCnt--; |
||
| 148 | } |
||
| 149 | |||
| 150 | /* If the numSamples is not a multiple of 4, compute any remaining output samples here. |
||
| 151 | ** No loop unrolling is used. */ |
||
| 152 | blkCnt = numSamples % 0x4u; |
||
| 153 | |||
| 154 | #else |
||
| 155 | |||
| 156 | /* Run the below code for Cortex-M0 */ |
||
| 157 | |||
| 158 | /* Initialize blkCnt with number of samples */ |
||
| 159 | blkCnt = numSamples; |
||
| 160 | |||
| 161 | #endif /* #ifndef ARM_MATH_CM0_FAMILY */ |
||
| 162 | |||
| 163 | while(blkCnt > 0u) |
||
| 164 | { |
||
| 165 | /* C(m,n) = A(m,n) * k */ |
||
| 166 | /* Scale, saturate and then store the results in the destination buffer. */ |
||
| 167 | *pOut++ = |
||
| 168 | (q15_t) (__SSAT(((q31_t) (*pIn++) * scaleFract) >> totShift, 16)); |
||
| 169 | |||
| 170 | /* Decrement the numSamples loop counter */ |
||
| 171 | blkCnt--; |
||
| 172 | } |
||
| 173 | /* Set status as ARM_MATH_SUCCESS */ |
||
| 174 | status = ARM_MATH_SUCCESS; |
||
| 175 | } |
||
| 176 | |||
| 177 | /* Return to application */ |
||
| 178 | return (status); |
||
| 179 | } |
||
| 180 | |||
| 181 | /** |
||
| 182 | * @} end of MatrixScale group |
||
| 183 | */ |