Eli Hughes / CMSIS_DSP_401

V4.0.1 of the ARM CMSIS DSP libraries. Note that arm_bitreversal2.s, arm_cfft_f32.c and arm_rfft_fast_f32.c had to be removed. arm_bitreversal2.s will not assemble with the online tools. So, the fast f32 FFT functions are not yet available. All the other FFT functions are available.

Dependents:   MPU9150_Example fir_f32 fir_f32 MPU9150_nucleo_noni2cdev ... more

FilteringFunctions/arm_conv_partial_opt_q15.c@0:3d9c67d97d6f, 2014-07-28 (annotated)

Committer:
emh203
Date:
Mon Jul 28 15:03:15 2014 +0000
Revision:
0:3d9c67d97d6f
1st working commit.   Had to remove arm_bitreversal2.s     arm_cfft_f32.c and arm_rfft_fast_f32.c.    The .s will not assemble.      For now I removed these functions so we could at least have a library for the other functions.

Who changed what in which revision?

UserRevisionLine numberNew contents of line
emh203 0:3d9c67d97d6f 1 /* ----------------------------------------------------------------------
emh203 0:3d9c67d97d6f 2 * Copyright (C) 2010-2014 ARM Limited. All rights reserved.
emh203 0:3d9c67d97d6f 3 *
emh203 0:3d9c67d97d6f 4 * $Date: 12. March 2014
emh203 0:3d9c67d97d6f 5 * $Revision: V1.4.3
emh203 0:3d9c67d97d6f 6 *
emh203 0:3d9c67d97d6f 7 * Project: CMSIS DSP Library
emh203 0:3d9c67d97d6f 8 * Title: arm_conv_partial_opt_q15.c
emh203 0:3d9c67d97d6f 9 *
emh203 0:3d9c67d97d6f 10 * Description: Partial convolution of Q15 sequences.
emh203 0:3d9c67d97d6f 11 *
emh203 0:3d9c67d97d6f 12 * Target Processor: Cortex-M4/Cortex-M3
emh203 0:3d9c67d97d6f 13 *
emh203 0:3d9c67d97d6f 14 * Redistribution and use in source and binary forms, with or without
emh203 0:3d9c67d97d6f 15 * modification, are permitted provided that the following conditions
emh203 0:3d9c67d97d6f 16 * are met:
emh203 0:3d9c67d97d6f 17 * - Redistributions of source code must retain the above copyright
emh203 0:3d9c67d97d6f 18 * notice, this list of conditions and the following disclaimer.
emh203 0:3d9c67d97d6f 19 * - Redistributions in binary form must reproduce the above copyright
emh203 0:3d9c67d97d6f 20 * notice, this list of conditions and the following disclaimer in
emh203 0:3d9c67d97d6f 21 * the documentation and/or other materials provided with the
emh203 0:3d9c67d97d6f 22 * distribution.
emh203 0:3d9c67d97d6f 23 * - Neither the name of ARM LIMITED nor the names of its contributors
emh203 0:3d9c67d97d6f 24 * may be used to endorse or promote products derived from this
emh203 0:3d9c67d97d6f 25 * software without specific prior written permission.
emh203 0:3d9c67d97d6f 26 *
emh203 0:3d9c67d97d6f 27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
emh203 0:3d9c67d97d6f 28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
emh203 0:3d9c67d97d6f 29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
emh203 0:3d9c67d97d6f 30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
emh203 0:3d9c67d97d6f 31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
emh203 0:3d9c67d97d6f 32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
emh203 0:3d9c67d97d6f 33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
emh203 0:3d9c67d97d6f 34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
emh203 0:3d9c67d97d6f 35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
emh203 0:3d9c67d97d6f 36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
emh203 0:3d9c67d97d6f 37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
emh203 0:3d9c67d97d6f 38 * POSSIBILITY OF SUCH DAMAGE.
emh203 0:3d9c67d97d6f 39 * -------------------------------------------------------------------- */
emh203 0:3d9c67d97d6f 40
emh203 0:3d9c67d97d6f 41 #include "arm_math.h"
emh203 0:3d9c67d97d6f 42
emh203 0:3d9c67d97d6f 43 /**
emh203 0:3d9c67d97d6f 44 * @ingroup groupFilters
emh203 0:3d9c67d97d6f 45 */
emh203 0:3d9c67d97d6f 46
emh203 0:3d9c67d97d6f 47 /**
emh203 0:3d9c67d97d6f 48 * @addtogroup PartialConv
emh203 0:3d9c67d97d6f 49 * @{
emh203 0:3d9c67d97d6f 50 */
emh203 0:3d9c67d97d6f 51
emh203 0:3d9c67d97d6f 52 /**
emh203 0:3d9c67d97d6f 53 * @brief Partial convolution of Q15 sequences.
emh203 0:3d9c67d97d6f 54 * @param[in] *pSrcA points to the first input sequence.
emh203 0:3d9c67d97d6f 55 * @param[in] srcALen length of the first input sequence.
emh203 0:3d9c67d97d6f 56 * @param[in] *pSrcB points to the second input sequence.
emh203 0:3d9c67d97d6f 57 * @param[in] srcBLen length of the second input sequence.
emh203 0:3d9c67d97d6f 58 * @param[out] *pDst points to the location where the output result is written.
emh203 0:3d9c67d97d6f 59 * @param[in] firstIndex is the first output sample to start with.
emh203 0:3d9c67d97d6f 60 * @param[in] numPoints is the number of output points to be computed.
emh203 0:3d9c67d97d6f 61 * @param[in] *pScratch1 points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
emh203 0:3d9c67d97d6f 62 * @param[in] *pScratch2 points to scratch buffer of size min(srcALen, srcBLen).
emh203 0:3d9c67d97d6f 63 * @return Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
emh203 0:3d9c67d97d6f 64 *
emh203 0:3d9c67d97d6f 65 * \par Restrictions
emh203 0:3d9c67d97d6f 66 * If the silicon does not support unaligned memory access enable the macro UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 67 * In this case input, output, state buffers should be aligned by 32-bit
emh203 0:3d9c67d97d6f 68 *
emh203 0:3d9c67d97d6f 69 * Refer to <code>arm_conv_partial_fast_q15()</code> for a faster but less precise version of this function for Cortex-M3 and Cortex-M4.
emh203 0:3d9c67d97d6f 70 *
emh203 0:3d9c67d97d6f 71 *
emh203 0:3d9c67d97d6f 72 */
emh203 0:3d9c67d97d6f 73
emh203 0:3d9c67d97d6f 74 #ifndef UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 75
emh203 0:3d9c67d97d6f 76 arm_status arm_conv_partial_opt_q15(
emh203 0:3d9c67d97d6f 77 q15_t * pSrcA,
emh203 0:3d9c67d97d6f 78 uint32_t srcALen,
emh203 0:3d9c67d97d6f 79 q15_t * pSrcB,
emh203 0:3d9c67d97d6f 80 uint32_t srcBLen,
emh203 0:3d9c67d97d6f 81 q15_t * pDst,
emh203 0:3d9c67d97d6f 82 uint32_t firstIndex,
emh203 0:3d9c67d97d6f 83 uint32_t numPoints,
emh203 0:3d9c67d97d6f 84 q15_t * pScratch1,
emh203 0:3d9c67d97d6f 85 q15_t * pScratch2)
emh203 0:3d9c67d97d6f 86 {
emh203 0:3d9c67d97d6f 87
emh203 0:3d9c67d97d6f 88 q15_t *pOut = pDst; /* output pointer */
emh203 0:3d9c67d97d6f 89 q15_t *pScr1 = pScratch1; /* Temporary pointer for scratch1 */
emh203 0:3d9c67d97d6f 90 q15_t *pScr2 = pScratch2; /* Temporary pointer for scratch1 */
emh203 0:3d9c67d97d6f 91 q63_t acc0, acc1, acc2, acc3; /* Accumulator */
emh203 0:3d9c67d97d6f 92 q31_t x1, x2, x3; /* Temporary variables to hold state and coefficient values */
emh203 0:3d9c67d97d6f 93 q31_t y1, y2; /* State variables */
emh203 0:3d9c67d97d6f 94 q15_t *pIn1; /* inputA pointer */
emh203 0:3d9c67d97d6f 95 q15_t *pIn2; /* inputB pointer */
emh203 0:3d9c67d97d6f 96 q15_t *px; /* Intermediate inputA pointer */
emh203 0:3d9c67d97d6f 97 q15_t *py; /* Intermediate inputB pointer */
emh203 0:3d9c67d97d6f 98 uint32_t j, k, blkCnt; /* loop counter */
emh203 0:3d9c67d97d6f 99 arm_status status; /* Status variable */
emh203 0:3d9c67d97d6f 100 uint32_t tapCnt; /* loop count */
emh203 0:3d9c67d97d6f 101
emh203 0:3d9c67d97d6f 102 /* Check for range of output samples to be calculated */
emh203 0:3d9c67d97d6f 103 if((firstIndex + numPoints) > ((srcALen + (srcBLen - 1u))))
emh203 0:3d9c67d97d6f 104 {
emh203 0:3d9c67d97d6f 105 /* Set status as ARM_MATH_ARGUMENT_ERROR */
emh203 0:3d9c67d97d6f 106 status = ARM_MATH_ARGUMENT_ERROR;
emh203 0:3d9c67d97d6f 107 }
emh203 0:3d9c67d97d6f 108 else
emh203 0:3d9c67d97d6f 109 {
emh203 0:3d9c67d97d6f 110
emh203 0:3d9c67d97d6f 111 /* The algorithm implementation is based on the lengths of the inputs. */
emh203 0:3d9c67d97d6f 112 /* srcB is always made to slide across srcA. */
emh203 0:3d9c67d97d6f 113 /* So srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 114 if(srcALen >= srcBLen)
emh203 0:3d9c67d97d6f 115 {
emh203 0:3d9c67d97d6f 116 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 117 pIn1 = pSrcA;
emh203 0:3d9c67d97d6f 118
emh203 0:3d9c67d97d6f 119 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 120 pIn2 = pSrcB;
emh203 0:3d9c67d97d6f 121 }
emh203 0:3d9c67d97d6f 122 else
emh203 0:3d9c67d97d6f 123 {
emh203 0:3d9c67d97d6f 124 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 125 pIn1 = pSrcB;
emh203 0:3d9c67d97d6f 126
emh203 0:3d9c67d97d6f 127 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 128 pIn2 = pSrcA;
emh203 0:3d9c67d97d6f 129
emh203 0:3d9c67d97d6f 130 /* srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 131 j = srcBLen;
emh203 0:3d9c67d97d6f 132 srcBLen = srcALen;
emh203 0:3d9c67d97d6f 133 srcALen = j;
emh203 0:3d9c67d97d6f 134 }
emh203 0:3d9c67d97d6f 135
emh203 0:3d9c67d97d6f 136 /* Temporary pointer for scratch2 */
emh203 0:3d9c67d97d6f 137 py = pScratch2;
emh203 0:3d9c67d97d6f 138
emh203 0:3d9c67d97d6f 139 /* pointer to take end of scratch2 buffer */
emh203 0:3d9c67d97d6f 140 pScr2 = pScratch2 + srcBLen - 1;
emh203 0:3d9c67d97d6f 141
emh203 0:3d9c67d97d6f 142 /* points to smaller length sequence */
emh203 0:3d9c67d97d6f 143 px = pIn2;
emh203 0:3d9c67d97d6f 144
emh203 0:3d9c67d97d6f 145 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 146 k = srcBLen >> 2u;
emh203 0:3d9c67d97d6f 147
emh203 0:3d9c67d97d6f 148 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 149 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 150 while(k > 0u)
emh203 0:3d9c67d97d6f 151 {
emh203 0:3d9c67d97d6f 152 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 153 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 154 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 155 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 156 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 157
emh203 0:3d9c67d97d6f 158 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 159 k--;
emh203 0:3d9c67d97d6f 160 }
emh203 0:3d9c67d97d6f 161
emh203 0:3d9c67d97d6f 162 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 163 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 164 k = srcBLen % 0x4u;
emh203 0:3d9c67d97d6f 165
emh203 0:3d9c67d97d6f 166 while(k > 0u)
emh203 0:3d9c67d97d6f 167 {
emh203 0:3d9c67d97d6f 168 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 169 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 170
emh203 0:3d9c67d97d6f 171 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 172 k--;
emh203 0:3d9c67d97d6f 173 }
emh203 0:3d9c67d97d6f 174
emh203 0:3d9c67d97d6f 175 /* Initialze temporary scratch pointer */
emh203 0:3d9c67d97d6f 176 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 177
emh203 0:3d9c67d97d6f 178 /* Fill (srcBLen - 1u) zeros in scratch buffer */
emh203 0:3d9c67d97d6f 179 arm_fill_q15(0, pScr1, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 180
emh203 0:3d9c67d97d6f 181 /* Update temporary scratch pointer */
emh203 0:3d9c67d97d6f 182 pScr1 += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 183
emh203 0:3d9c67d97d6f 184 /* Copy bigger length sequence(srcALen) samples in scratch1 buffer */
emh203 0:3d9c67d97d6f 185
emh203 0:3d9c67d97d6f 186 /* Copy (srcALen) samples in scratch buffer */
emh203 0:3d9c67d97d6f 187 arm_copy_q15(pIn1, pScr1, srcALen);
emh203 0:3d9c67d97d6f 188
emh203 0:3d9c67d97d6f 189 /* Update pointers */
emh203 0:3d9c67d97d6f 190 pScr1 += srcALen;
emh203 0:3d9c67d97d6f 191
emh203 0:3d9c67d97d6f 192 /* Fill (srcBLen - 1u) zeros at end of scratch buffer */
emh203 0:3d9c67d97d6f 193 arm_fill_q15(0, pScr1, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 194
emh203 0:3d9c67d97d6f 195 /* Update pointer */
emh203 0:3d9c67d97d6f 196 pScr1 += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 197
emh203 0:3d9c67d97d6f 198 /* Initialization of pIn2 pointer */
emh203 0:3d9c67d97d6f 199 pIn2 = py;
emh203 0:3d9c67d97d6f 200
emh203 0:3d9c67d97d6f 201 pScratch1 += firstIndex;
emh203 0:3d9c67d97d6f 202
emh203 0:3d9c67d97d6f 203 pOut = pDst + firstIndex;
emh203 0:3d9c67d97d6f 204
emh203 0:3d9c67d97d6f 205 /* Actual convolution process starts here */
emh203 0:3d9c67d97d6f 206 blkCnt = (numPoints) >> 2;
emh203 0:3d9c67d97d6f 207
emh203 0:3d9c67d97d6f 208 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 209 {
emh203 0:3d9c67d97d6f 210 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 211 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 212
emh203 0:3d9c67d97d6f 213 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 214 acc0 = 0;
emh203 0:3d9c67d97d6f 215 acc1 = 0;
emh203 0:3d9c67d97d6f 216 acc2 = 0;
emh203 0:3d9c67d97d6f 217 acc3 = 0;
emh203 0:3d9c67d97d6f 218
emh203 0:3d9c67d97d6f 219 /* Read two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 220 x1 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 221
emh203 0:3d9c67d97d6f 222 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 223 x2 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 224
emh203 0:3d9c67d97d6f 225 tapCnt = (srcBLen) >> 2u;
emh203 0:3d9c67d97d6f 226
emh203 0:3d9c67d97d6f 227 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 228 {
emh203 0:3d9c67d97d6f 229
emh203 0:3d9c67d97d6f 230 /* Read four samples from smaller buffer */
emh203 0:3d9c67d97d6f 231 y1 = _SIMD32_OFFSET(pIn2);
emh203 0:3d9c67d97d6f 232 y2 = _SIMD32_OFFSET(pIn2 + 2u);
emh203 0:3d9c67d97d6f 233
emh203 0:3d9c67d97d6f 234 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 235 acc0 = __SMLALD(x1, y1, acc0);
emh203 0:3d9c67d97d6f 236 acc2 = __SMLALD(x2, y1, acc2);
emh203 0:3d9c67d97d6f 237
emh203 0:3d9c67d97d6f 238 /* pack input data */
emh203 0:3d9c67d97d6f 239 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 240 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 241 #else
emh203 0:3d9c67d97d6f 242 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 243 #endif
emh203 0:3d9c67d97d6f 244
emh203 0:3d9c67d97d6f 245 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 246 acc1 = __SMLALDX(x3, y1, acc1);
emh203 0:3d9c67d97d6f 247
emh203 0:3d9c67d97d6f 248 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 249 x1 = _SIMD32_OFFSET(pScr1);
emh203 0:3d9c67d97d6f 250
emh203 0:3d9c67d97d6f 251 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 252 acc0 = __SMLALD(x2, y2, acc0);
emh203 0:3d9c67d97d6f 253 acc2 = __SMLALD(x1, y2, acc2);
emh203 0:3d9c67d97d6f 254
emh203 0:3d9c67d97d6f 255 /* pack input data */
emh203 0:3d9c67d97d6f 256 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 257 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 258 #else
emh203 0:3d9c67d97d6f 259 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 260 #endif
emh203 0:3d9c67d97d6f 261
emh203 0:3d9c67d97d6f 262 acc3 = __SMLALDX(x3, y1, acc3);
emh203 0:3d9c67d97d6f 263 acc1 = __SMLALDX(x3, y2, acc1);
emh203 0:3d9c67d97d6f 264
emh203 0:3d9c67d97d6f 265 x2 = _SIMD32_OFFSET(pScr1 + 2u);
emh203 0:3d9c67d97d6f 266
emh203 0:3d9c67d97d6f 267 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 268 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 269 #else
emh203 0:3d9c67d97d6f 270 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 271 #endif
emh203 0:3d9c67d97d6f 272
emh203 0:3d9c67d97d6f 273 acc3 = __SMLALDX(x3, y2, acc3);
emh203 0:3d9c67d97d6f 274
emh203 0:3d9c67d97d6f 275 /* update scratch pointers */
emh203 0:3d9c67d97d6f 276 pIn2 += 4u;
emh203 0:3d9c67d97d6f 277 pScr1 += 4u;
emh203 0:3d9c67d97d6f 278
emh203 0:3d9c67d97d6f 279
emh203 0:3d9c67d97d6f 280 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 281 tapCnt--;
emh203 0:3d9c67d97d6f 282 }
emh203 0:3d9c67d97d6f 283
emh203 0:3d9c67d97d6f 284 /* Update scratch pointer for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 285 pScr1 -= 4u;
emh203 0:3d9c67d97d6f 286
emh203 0:3d9c67d97d6f 287 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 288 tapCnt = (srcBLen) & 3u;
emh203 0:3d9c67d97d6f 289
emh203 0:3d9c67d97d6f 290 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 291 {
emh203 0:3d9c67d97d6f 292 /* accumlate the results */
emh203 0:3d9c67d97d6f 293 acc0 += (*pScr1++ * *pIn2);
emh203 0:3d9c67d97d6f 294 acc1 += (*pScr1++ * *pIn2);
emh203 0:3d9c67d97d6f 295 acc2 += (*pScr1++ * *pIn2);
emh203 0:3d9c67d97d6f 296 acc3 += (*pScr1++ * *pIn2++);
emh203 0:3d9c67d97d6f 297
emh203 0:3d9c67d97d6f 298 pScr1 -= 3u;
emh203 0:3d9c67d97d6f 299
emh203 0:3d9c67d97d6f 300 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 301 tapCnt--;
emh203 0:3d9c67d97d6f 302 }
emh203 0:3d9c67d97d6f 303
emh203 0:3d9c67d97d6f 304 blkCnt--;
emh203 0:3d9c67d97d6f 305
emh203 0:3d9c67d97d6f 306
emh203 0:3d9c67d97d6f 307 /* Store the results in the accumulators in the destination buffer. */
emh203 0:3d9c67d97d6f 308
emh203 0:3d9c67d97d6f 309 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 310
emh203 0:3d9c67d97d6f 311 *__SIMD32(pOut)++ =
emh203 0:3d9c67d97d6f 312 __PKHBT(__SSAT((acc0 >> 15), 16), __SSAT((acc1 >> 15), 16), 16);
emh203 0:3d9c67d97d6f 313 *__SIMD32(pOut)++ =
emh203 0:3d9c67d97d6f 314 __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16);
emh203 0:3d9c67d97d6f 315
emh203 0:3d9c67d97d6f 316 #else
emh203 0:3d9c67d97d6f 317
emh203 0:3d9c67d97d6f 318 *__SIMD32(pOut)++ =
emh203 0:3d9c67d97d6f 319 __PKHBT(__SSAT((acc1 >> 15), 16), __SSAT((acc0 >> 15), 16), 16);
emh203 0:3d9c67d97d6f 320 *__SIMD32(pOut)++ =
emh203 0:3d9c67d97d6f 321 __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16);
emh203 0:3d9c67d97d6f 322
emh203 0:3d9c67d97d6f 323 #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
emh203 0:3d9c67d97d6f 324
emh203 0:3d9c67d97d6f 325 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 326 pIn2 = py;
emh203 0:3d9c67d97d6f 327
emh203 0:3d9c67d97d6f 328 pScratch1 += 4u;
emh203 0:3d9c67d97d6f 329
emh203 0:3d9c67d97d6f 330 }
emh203 0:3d9c67d97d6f 331
emh203 0:3d9c67d97d6f 332
emh203 0:3d9c67d97d6f 333 blkCnt = numPoints & 0x3;
emh203 0:3d9c67d97d6f 334
emh203 0:3d9c67d97d6f 335 /* Calculate convolution for remaining samples of Bigger length sequence */
emh203 0:3d9c67d97d6f 336 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 337 {
emh203 0:3d9c67d97d6f 338 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 339 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 340
emh203 0:3d9c67d97d6f 341 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 342 acc0 = 0;
emh203 0:3d9c67d97d6f 343
emh203 0:3d9c67d97d6f 344 tapCnt = (srcBLen) >> 1u;
emh203 0:3d9c67d97d6f 345
emh203 0:3d9c67d97d6f 346 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 347 {
emh203 0:3d9c67d97d6f 348
emh203 0:3d9c67d97d6f 349 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 350 x1 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 351
emh203 0:3d9c67d97d6f 352 /* Read two samples from smaller buffer */
emh203 0:3d9c67d97d6f 353 y1 = *__SIMD32(pIn2)++;
emh203 0:3d9c67d97d6f 354
emh203 0:3d9c67d97d6f 355 acc0 = __SMLALD(x1, y1, acc0);
emh203 0:3d9c67d97d6f 356
emh203 0:3d9c67d97d6f 357 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 358 tapCnt--;
emh203 0:3d9c67d97d6f 359 }
emh203 0:3d9c67d97d6f 360
emh203 0:3d9c67d97d6f 361 tapCnt = (srcBLen) & 1u;
emh203 0:3d9c67d97d6f 362
emh203 0:3d9c67d97d6f 363 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 364 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 365 {
emh203 0:3d9c67d97d6f 366
emh203 0:3d9c67d97d6f 367 /* accumlate the results */
emh203 0:3d9c67d97d6f 368 acc0 += (*pScr1++ * *pIn2++);
emh203 0:3d9c67d97d6f 369
emh203 0:3d9c67d97d6f 370 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 371 tapCnt--;
emh203 0:3d9c67d97d6f 372 }
emh203 0:3d9c67d97d6f 373
emh203 0:3d9c67d97d6f 374 blkCnt--;
emh203 0:3d9c67d97d6f 375
emh203 0:3d9c67d97d6f 376 /* Store the result in the accumulator in the destination buffer. */
emh203 0:3d9c67d97d6f 377 *pOut++ = (q15_t) (__SSAT((acc0 >> 15), 16));
emh203 0:3d9c67d97d6f 378
emh203 0:3d9c67d97d6f 379 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 380 pIn2 = py;
emh203 0:3d9c67d97d6f 381
emh203 0:3d9c67d97d6f 382 pScratch1 += 1u;
emh203 0:3d9c67d97d6f 383
emh203 0:3d9c67d97d6f 384 }
emh203 0:3d9c67d97d6f 385
emh203 0:3d9c67d97d6f 386 /* set status as ARM_MATH_SUCCESS */
emh203 0:3d9c67d97d6f 387 status = ARM_MATH_SUCCESS;
emh203 0:3d9c67d97d6f 388
emh203 0:3d9c67d97d6f 389 }
emh203 0:3d9c67d97d6f 390
emh203 0:3d9c67d97d6f 391 /* Return to application */
emh203 0:3d9c67d97d6f 392 return (status);
emh203 0:3d9c67d97d6f 393 }
emh203 0:3d9c67d97d6f 394
emh203 0:3d9c67d97d6f 395 #else
emh203 0:3d9c67d97d6f 396
emh203 0:3d9c67d97d6f 397 arm_status arm_conv_partial_opt_q15(
emh203 0:3d9c67d97d6f 398 q15_t * pSrcA,
emh203 0:3d9c67d97d6f 399 uint32_t srcALen,
emh203 0:3d9c67d97d6f 400 q15_t * pSrcB,
emh203 0:3d9c67d97d6f 401 uint32_t srcBLen,
emh203 0:3d9c67d97d6f 402 q15_t * pDst,
emh203 0:3d9c67d97d6f 403 uint32_t firstIndex,
emh203 0:3d9c67d97d6f 404 uint32_t numPoints,
emh203 0:3d9c67d97d6f 405 q15_t * pScratch1,
emh203 0:3d9c67d97d6f 406 q15_t * pScratch2)
emh203 0:3d9c67d97d6f 407 {
emh203 0:3d9c67d97d6f 408
emh203 0:3d9c67d97d6f 409 q15_t *pOut = pDst; /* output pointer */
emh203 0:3d9c67d97d6f 410 q15_t *pScr1 = pScratch1; /* Temporary pointer for scratch1 */
emh203 0:3d9c67d97d6f 411 q15_t *pScr2 = pScratch2; /* Temporary pointer for scratch1 */
emh203 0:3d9c67d97d6f 412 q63_t acc0, acc1, acc2, acc3; /* Accumulator */
emh203 0:3d9c67d97d6f 413 q15_t *pIn1; /* inputA pointer */
emh203 0:3d9c67d97d6f 414 q15_t *pIn2; /* inputB pointer */
emh203 0:3d9c67d97d6f 415 q15_t *px; /* Intermediate inputA pointer */
emh203 0:3d9c67d97d6f 416 q15_t *py; /* Intermediate inputB pointer */
emh203 0:3d9c67d97d6f 417 uint32_t j, k, blkCnt; /* loop counter */
emh203 0:3d9c67d97d6f 418 arm_status status; /* Status variable */
emh203 0:3d9c67d97d6f 419 uint32_t tapCnt; /* loop count */
emh203 0:3d9c67d97d6f 420 q15_t x10, x11, x20, x21; /* Temporary variables to hold srcA buffer */
emh203 0:3d9c67d97d6f 421 q15_t y10, y11; /* Temporary variables to hold srcB buffer */
emh203 0:3d9c67d97d6f 422
emh203 0:3d9c67d97d6f 423
emh203 0:3d9c67d97d6f 424 /* Check for range of output samples to be calculated */
emh203 0:3d9c67d97d6f 425 if((firstIndex + numPoints) > ((srcALen + (srcBLen - 1u))))
emh203 0:3d9c67d97d6f 426 {
emh203 0:3d9c67d97d6f 427 /* Set status as ARM_MATH_ARGUMENT_ERROR */
emh203 0:3d9c67d97d6f 428 status = ARM_MATH_ARGUMENT_ERROR;
emh203 0:3d9c67d97d6f 429 }
emh203 0:3d9c67d97d6f 430 else
emh203 0:3d9c67d97d6f 431 {
emh203 0:3d9c67d97d6f 432
emh203 0:3d9c67d97d6f 433 /* The algorithm implementation is based on the lengths of the inputs. */
emh203 0:3d9c67d97d6f 434 /* srcB is always made to slide across srcA. */
emh203 0:3d9c67d97d6f 435 /* So srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 436 if(srcALen >= srcBLen)
emh203 0:3d9c67d97d6f 437 {
emh203 0:3d9c67d97d6f 438 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 439 pIn1 = pSrcA;
emh203 0:3d9c67d97d6f 440
emh203 0:3d9c67d97d6f 441 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 442 pIn2 = pSrcB;
emh203 0:3d9c67d97d6f 443 }
emh203 0:3d9c67d97d6f 444 else
emh203 0:3d9c67d97d6f 445 {
emh203 0:3d9c67d97d6f 446 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 447 pIn1 = pSrcB;
emh203 0:3d9c67d97d6f 448
emh203 0:3d9c67d97d6f 449 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 450 pIn2 = pSrcA;
emh203 0:3d9c67d97d6f 451
emh203 0:3d9c67d97d6f 452 /* srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 453 j = srcBLen;
emh203 0:3d9c67d97d6f 454 srcBLen = srcALen;
emh203 0:3d9c67d97d6f 455 srcALen = j;
emh203 0:3d9c67d97d6f 456 }
emh203 0:3d9c67d97d6f 457
emh203 0:3d9c67d97d6f 458 /* Temporary pointer for scratch2 */
emh203 0:3d9c67d97d6f 459 py = pScratch2;
emh203 0:3d9c67d97d6f 460
emh203 0:3d9c67d97d6f 461 /* pointer to take end of scratch2 buffer */
emh203 0:3d9c67d97d6f 462 pScr2 = pScratch2 + srcBLen - 1;
emh203 0:3d9c67d97d6f 463
emh203 0:3d9c67d97d6f 464 /* points to smaller length sequence */
emh203 0:3d9c67d97d6f 465 px = pIn2;
emh203 0:3d9c67d97d6f 466
emh203 0:3d9c67d97d6f 467 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 468 k = srcBLen >> 2u;
emh203 0:3d9c67d97d6f 469
emh203 0:3d9c67d97d6f 470 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 471 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 472 while(k > 0u)
emh203 0:3d9c67d97d6f 473 {
emh203 0:3d9c67d97d6f 474 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 475 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 476 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 477 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 478 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 479
emh203 0:3d9c67d97d6f 480 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 481 k--;
emh203 0:3d9c67d97d6f 482 }
emh203 0:3d9c67d97d6f 483
emh203 0:3d9c67d97d6f 484 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 485 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 486 k = srcBLen % 0x4u;
emh203 0:3d9c67d97d6f 487
emh203 0:3d9c67d97d6f 488 while(k > 0u)
emh203 0:3d9c67d97d6f 489 {
emh203 0:3d9c67d97d6f 490 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 491 *pScr2-- = *px++;
emh203 0:3d9c67d97d6f 492
emh203 0:3d9c67d97d6f 493 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 494 k--;
emh203 0:3d9c67d97d6f 495 }
emh203 0:3d9c67d97d6f 496
emh203 0:3d9c67d97d6f 497 /* Initialze temporary scratch pointer */
emh203 0:3d9c67d97d6f 498 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 499
emh203 0:3d9c67d97d6f 500 /* Fill (srcBLen - 1u) zeros in scratch buffer */
emh203 0:3d9c67d97d6f 501 arm_fill_q15(0, pScr1, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 502
emh203 0:3d9c67d97d6f 503 /* Update temporary scratch pointer */
emh203 0:3d9c67d97d6f 504 pScr1 += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 505
emh203 0:3d9c67d97d6f 506 /* Copy bigger length sequence(srcALen) samples in scratch1 buffer */
emh203 0:3d9c67d97d6f 507
emh203 0:3d9c67d97d6f 508
emh203 0:3d9c67d97d6f 509 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 510 k = srcALen >> 2u;
emh203 0:3d9c67d97d6f 511
emh203 0:3d9c67d97d6f 512 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 513 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 514 while(k > 0u)
emh203 0:3d9c67d97d6f 515 {
emh203 0:3d9c67d97d6f 516 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 517 *pScr1++ = *pIn1++;
emh203 0:3d9c67d97d6f 518 *pScr1++ = *pIn1++;
emh203 0:3d9c67d97d6f 519 *pScr1++ = *pIn1++;
emh203 0:3d9c67d97d6f 520 *pScr1++ = *pIn1++;
emh203 0:3d9c67d97d6f 521
emh203 0:3d9c67d97d6f 522 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 523 k--;
emh203 0:3d9c67d97d6f 524 }
emh203 0:3d9c67d97d6f 525
emh203 0:3d9c67d97d6f 526 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 527 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 528 k = srcALen % 0x4u;
emh203 0:3d9c67d97d6f 529
emh203 0:3d9c67d97d6f 530 while(k > 0u)
emh203 0:3d9c67d97d6f 531 {
emh203 0:3d9c67d97d6f 532 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 533 *pScr1++ = *pIn1++;
emh203 0:3d9c67d97d6f 534
emh203 0:3d9c67d97d6f 535 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 536 k--;
emh203 0:3d9c67d97d6f 537 }
emh203 0:3d9c67d97d6f 538
emh203 0:3d9c67d97d6f 539
emh203 0:3d9c67d97d6f 540 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 541 k = (srcBLen - 1u) >> 2u;
emh203 0:3d9c67d97d6f 542
emh203 0:3d9c67d97d6f 543 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 544 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 545 while(k > 0u)
emh203 0:3d9c67d97d6f 546 {
emh203 0:3d9c67d97d6f 547 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 548 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 549 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 550 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 551 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 552
emh203 0:3d9c67d97d6f 553 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 554 k--;
emh203 0:3d9c67d97d6f 555 }
emh203 0:3d9c67d97d6f 556
emh203 0:3d9c67d97d6f 557 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 558 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 559 k = (srcBLen - 1u) % 0x4u;
emh203 0:3d9c67d97d6f 560
emh203 0:3d9c67d97d6f 561 while(k > 0u)
emh203 0:3d9c67d97d6f 562 {
emh203 0:3d9c67d97d6f 563 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 564 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 565
emh203 0:3d9c67d97d6f 566 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 567 k--;
emh203 0:3d9c67d97d6f 568 }
emh203 0:3d9c67d97d6f 569
emh203 0:3d9c67d97d6f 570
emh203 0:3d9c67d97d6f 571 /* Initialization of pIn2 pointer */
emh203 0:3d9c67d97d6f 572 pIn2 = py;
emh203 0:3d9c67d97d6f 573
emh203 0:3d9c67d97d6f 574 pScratch1 += firstIndex;
emh203 0:3d9c67d97d6f 575
emh203 0:3d9c67d97d6f 576 pOut = pDst + firstIndex;
emh203 0:3d9c67d97d6f 577
emh203 0:3d9c67d97d6f 578 /* Actual convolution process starts here */
emh203 0:3d9c67d97d6f 579 blkCnt = (numPoints) >> 2;
emh203 0:3d9c67d97d6f 580
emh203 0:3d9c67d97d6f 581 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 582 {
emh203 0:3d9c67d97d6f 583 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 584 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 585
emh203 0:3d9c67d97d6f 586 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 587 acc0 = 0;
emh203 0:3d9c67d97d6f 588 acc1 = 0;
emh203 0:3d9c67d97d6f 589 acc2 = 0;
emh203 0:3d9c67d97d6f 590 acc3 = 0;
emh203 0:3d9c67d97d6f 591
emh203 0:3d9c67d97d6f 592 /* Read two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 593 x10 = *pScr1++;
emh203 0:3d9c67d97d6f 594 x11 = *pScr1++;
emh203 0:3d9c67d97d6f 595
emh203 0:3d9c67d97d6f 596 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 597 x20 = *pScr1++;
emh203 0:3d9c67d97d6f 598 x21 = *pScr1++;
emh203 0:3d9c67d97d6f 599
emh203 0:3d9c67d97d6f 600 tapCnt = (srcBLen) >> 2u;
emh203 0:3d9c67d97d6f 601
emh203 0:3d9c67d97d6f 602 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 603 {
emh203 0:3d9c67d97d6f 604
emh203 0:3d9c67d97d6f 605 /* Read two samples from smaller buffer */
emh203 0:3d9c67d97d6f 606 y10 = *pIn2;
emh203 0:3d9c67d97d6f 607 y11 = *(pIn2 + 1u);
emh203 0:3d9c67d97d6f 608
emh203 0:3d9c67d97d6f 609 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 610 acc0 += (q63_t) x10 *y10;
emh203 0:3d9c67d97d6f 611 acc0 += (q63_t) x11 *y11;
emh203 0:3d9c67d97d6f 612 acc2 += (q63_t) x20 *y10;
emh203 0:3d9c67d97d6f 613 acc2 += (q63_t) x21 *y11;
emh203 0:3d9c67d97d6f 614
emh203 0:3d9c67d97d6f 615 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 616 acc1 += (q63_t) x11 *y10;
emh203 0:3d9c67d97d6f 617 acc1 += (q63_t) x20 *y11;
emh203 0:3d9c67d97d6f 618
emh203 0:3d9c67d97d6f 619 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 620 x10 = *pScr1;
emh203 0:3d9c67d97d6f 621 x11 = *(pScr1 + 1u);
emh203 0:3d9c67d97d6f 622
emh203 0:3d9c67d97d6f 623 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 624 acc3 += (q63_t) x21 *y10;
emh203 0:3d9c67d97d6f 625 acc3 += (q63_t) x10 *y11;
emh203 0:3d9c67d97d6f 626
emh203 0:3d9c67d97d6f 627 /* Read next two samples from scratch2 buffer */
emh203 0:3d9c67d97d6f 628 y10 = *(pIn2 + 2u);
emh203 0:3d9c67d97d6f 629 y11 = *(pIn2 + 3u);
emh203 0:3d9c67d97d6f 630
emh203 0:3d9c67d97d6f 631 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 632 acc0 += (q63_t) x20 *y10;
emh203 0:3d9c67d97d6f 633 acc0 += (q63_t) x21 *y11;
emh203 0:3d9c67d97d6f 634 acc2 += (q63_t) x10 *y10;
emh203 0:3d9c67d97d6f 635 acc2 += (q63_t) x11 *y11;
emh203 0:3d9c67d97d6f 636 acc1 += (q63_t) x21 *y10;
emh203 0:3d9c67d97d6f 637 acc1 += (q63_t) x10 *y11;
emh203 0:3d9c67d97d6f 638
emh203 0:3d9c67d97d6f 639 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 640 x20 = *(pScr1 + 2);
emh203 0:3d9c67d97d6f 641 x21 = *(pScr1 + 3);
emh203 0:3d9c67d97d6f 642
emh203 0:3d9c67d97d6f 643 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 644 acc3 += (q63_t) x11 *y10;
emh203 0:3d9c67d97d6f 645 acc3 += (q63_t) x20 *y11;
emh203 0:3d9c67d97d6f 646
emh203 0:3d9c67d97d6f 647 /* update scratch pointers */
emh203 0:3d9c67d97d6f 648 pIn2 += 4u;
emh203 0:3d9c67d97d6f 649 pScr1 += 4u;
emh203 0:3d9c67d97d6f 650
emh203 0:3d9c67d97d6f 651 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 652 tapCnt--;
emh203 0:3d9c67d97d6f 653 }
emh203 0:3d9c67d97d6f 654
emh203 0:3d9c67d97d6f 655 /* Update scratch pointer for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 656 pScr1 -= 4u;
emh203 0:3d9c67d97d6f 657
emh203 0:3d9c67d97d6f 658 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 659 tapCnt = (srcBLen) & 3u;
emh203 0:3d9c67d97d6f 660
emh203 0:3d9c67d97d6f 661 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 662 {
emh203 0:3d9c67d97d6f 663 /* accumlate the results */
emh203 0:3d9c67d97d6f 664 acc0 += (*pScr1++ * *pIn2);
emh203 0:3d9c67d97d6f 665 acc1 += (*pScr1++ * *pIn2);
emh203 0:3d9c67d97d6f 666 acc2 += (*pScr1++ * *pIn2);
emh203 0:3d9c67d97d6f 667 acc3 += (*pScr1++ * *pIn2++);
emh203 0:3d9c67d97d6f 668
emh203 0:3d9c67d97d6f 669 pScr1 -= 3u;
emh203 0:3d9c67d97d6f 670
emh203 0:3d9c67d97d6f 671 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 672 tapCnt--;
emh203 0:3d9c67d97d6f 673 }
emh203 0:3d9c67d97d6f 674
emh203 0:3d9c67d97d6f 675 blkCnt--;
emh203 0:3d9c67d97d6f 676
emh203 0:3d9c67d97d6f 677
emh203 0:3d9c67d97d6f 678 /* Store the results in the accumulators in the destination buffer. */
emh203 0:3d9c67d97d6f 679 *pOut++ = __SSAT((acc0 >> 15), 16);
emh203 0:3d9c67d97d6f 680 *pOut++ = __SSAT((acc1 >> 15), 16);
emh203 0:3d9c67d97d6f 681 *pOut++ = __SSAT((acc2 >> 15), 16);
emh203 0:3d9c67d97d6f 682 *pOut++ = __SSAT((acc3 >> 15), 16);
emh203 0:3d9c67d97d6f 683
emh203 0:3d9c67d97d6f 684
emh203 0:3d9c67d97d6f 685 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 686 pIn2 = py;
emh203 0:3d9c67d97d6f 687
emh203 0:3d9c67d97d6f 688 pScratch1 += 4u;
emh203 0:3d9c67d97d6f 689
emh203 0:3d9c67d97d6f 690 }
emh203 0:3d9c67d97d6f 691
emh203 0:3d9c67d97d6f 692
emh203 0:3d9c67d97d6f 693 blkCnt = numPoints & 0x3;
emh203 0:3d9c67d97d6f 694
emh203 0:3d9c67d97d6f 695 /* Calculate convolution for remaining samples of Bigger length sequence */
emh203 0:3d9c67d97d6f 696 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 697 {
emh203 0:3d9c67d97d6f 698 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 699 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 700
emh203 0:3d9c67d97d6f 701 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 702 acc0 = 0;
emh203 0:3d9c67d97d6f 703
emh203 0:3d9c67d97d6f 704 tapCnt = (srcBLen) >> 1u;
emh203 0:3d9c67d97d6f 705
emh203 0:3d9c67d97d6f 706 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 707 {
emh203 0:3d9c67d97d6f 708
emh203 0:3d9c67d97d6f 709 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 710 x10 = *pScr1++;
emh203 0:3d9c67d97d6f 711 x11 = *pScr1++;
emh203 0:3d9c67d97d6f 712
emh203 0:3d9c67d97d6f 713 /* Read two samples from smaller buffer */
emh203 0:3d9c67d97d6f 714 y10 = *pIn2++;
emh203 0:3d9c67d97d6f 715 y11 = *pIn2++;
emh203 0:3d9c67d97d6f 716
emh203 0:3d9c67d97d6f 717 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 718 acc0 += (q63_t) x10 *y10;
emh203 0:3d9c67d97d6f 719 acc0 += (q63_t) x11 *y11;
emh203 0:3d9c67d97d6f 720
emh203 0:3d9c67d97d6f 721 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 722 tapCnt--;
emh203 0:3d9c67d97d6f 723 }
emh203 0:3d9c67d97d6f 724
emh203 0:3d9c67d97d6f 725 tapCnt = (srcBLen) & 1u;
emh203 0:3d9c67d97d6f 726
emh203 0:3d9c67d97d6f 727 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 728 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 729 {
emh203 0:3d9c67d97d6f 730
emh203 0:3d9c67d97d6f 731 /* accumlate the results */
emh203 0:3d9c67d97d6f 732 acc0 += (*pScr1++ * *pIn2++);
emh203 0:3d9c67d97d6f 733
emh203 0:3d9c67d97d6f 734 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 735 tapCnt--;
emh203 0:3d9c67d97d6f 736 }
emh203 0:3d9c67d97d6f 737
emh203 0:3d9c67d97d6f 738 blkCnt--;
emh203 0:3d9c67d97d6f 739
emh203 0:3d9c67d97d6f 740 /* Store the result in the accumulator in the destination buffer. */
emh203 0:3d9c67d97d6f 741 *pOut++ = (q15_t) (__SSAT((acc0 >> 15), 16));
emh203 0:3d9c67d97d6f 742
emh203 0:3d9c67d97d6f 743
emh203 0:3d9c67d97d6f 744 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 745 pIn2 = py;
emh203 0:3d9c67d97d6f 746
emh203 0:3d9c67d97d6f 747 pScratch1 += 1u;
emh203 0:3d9c67d97d6f 748
emh203 0:3d9c67d97d6f 749 }
emh203 0:3d9c67d97d6f 750
emh203 0:3d9c67d97d6f 751 /* set status as ARM_MATH_SUCCESS */
emh203 0:3d9c67d97d6f 752 status = ARM_MATH_SUCCESS;
emh203 0:3d9c67d97d6f 753
emh203 0:3d9c67d97d6f 754 }
emh203 0:3d9c67d97d6f 755
emh203 0:3d9c67d97d6f 756 /* Return to application */
emh203 0:3d9c67d97d6f 757 return (status);
emh203 0:3d9c67d97d6f 758 }
emh203 0:3d9c67d97d6f 759
emh203 0:3d9c67d97d6f 760 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emh203 0:3d9c67d97d6f 761
emh203 0:3d9c67d97d6f 762
emh203 0:3d9c67d97d6f 763 /**
emh203 0:3d9c67d97d6f 764 * @} end of PartialConv group
emh203 0:3d9c67d97d6f 765 */