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_correlate_fast_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_correlate_fast_opt_q15.c
emh203 0:3d9c67d97d6f 9 *
emh203 0:3d9c67d97d6f 10 * Description: Fast Q15 Correlation.
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 Corr
emh203 0:3d9c67d97d6f 49 * @{
emh203 0:3d9c67d97d6f 50 */
emh203 0:3d9c67d97d6f 51
emh203 0:3d9c67d97d6f 52 /**
emh203 0:3d9c67d97d6f 53 * @brief Correlation of Q15 sequences (fast version) for Cortex-M3 and Cortex-M4.
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. Length 2 * max(srcALen, srcBLen) - 1.
emh203 0:3d9c67d97d6f 59 * @param[in] *pScratch points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
emh203 0:3d9c67d97d6f 60 * @return none.
emh203 0:3d9c67d97d6f 61 *
emh203 0:3d9c67d97d6f 62 *
emh203 0:3d9c67d97d6f 63 * \par Restrictions
emh203 0:3d9c67d97d6f 64 * If the silicon does not support unaligned memory access enable the macro UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 65 * In this case input, output, scratch buffers should be aligned by 32-bit
emh203 0:3d9c67d97d6f 66 *
emh203 0:3d9c67d97d6f 67 *
emh203 0:3d9c67d97d6f 68 * <b>Scaling and Overflow Behavior:</b>
emh203 0:3d9c67d97d6f 69 *
emh203 0:3d9c67d97d6f 70 * \par
emh203 0:3d9c67d97d6f 71 * This fast version uses a 32-bit accumulator with 2.30 format.
emh203 0:3d9c67d97d6f 72 * The accumulator maintains full precision of the intermediate multiplication results but provides only a single guard bit.
emh203 0:3d9c67d97d6f 73 * There is no saturation on intermediate additions.
emh203 0:3d9c67d97d6f 74 * Thus, if the accumulator overflows it wraps around and distorts the result.
emh203 0:3d9c67d97d6f 75 * The input signals should be scaled down to avoid intermediate overflows.
emh203 0:3d9c67d97d6f 76 * Scale down one of the inputs by 1/min(srcALen, srcBLen) to avoid overflow since a
emh203 0:3d9c67d97d6f 77 * maximum of min(srcALen, srcBLen) number of additions is carried internally.
emh203 0:3d9c67d97d6f 78 * The 2.30 accumulator is right shifted by 15 bits and then saturated to 1.15 format to yield the final result.
emh203 0:3d9c67d97d6f 79 *
emh203 0:3d9c67d97d6f 80 * \par
emh203 0:3d9c67d97d6f 81 * See <code>arm_correlate_q15()</code> for a slower implementation of this function which uses a 64-bit accumulator to avoid wrap around distortion.
emh203 0:3d9c67d97d6f 82 */
emh203 0:3d9c67d97d6f 83
emh203 0:3d9c67d97d6f 84 void arm_correlate_fast_opt_q15(
emh203 0:3d9c67d97d6f 85 q15_t * pSrcA,
emh203 0:3d9c67d97d6f 86 uint32_t srcALen,
emh203 0:3d9c67d97d6f 87 q15_t * pSrcB,
emh203 0:3d9c67d97d6f 88 uint32_t srcBLen,
emh203 0:3d9c67d97d6f 89 q15_t * pDst,
emh203 0:3d9c67d97d6f 90 q15_t * pScratch)
emh203 0:3d9c67d97d6f 91 {
emh203 0:3d9c67d97d6f 92 q15_t *pIn1; /* inputA pointer */
emh203 0:3d9c67d97d6f 93 q15_t *pIn2; /* inputB pointer */
emh203 0:3d9c67d97d6f 94 q31_t acc0, acc1, acc2, acc3; /* Accumulators */
emh203 0:3d9c67d97d6f 95 q15_t *py; /* Intermediate inputB pointer */
emh203 0:3d9c67d97d6f 96 q31_t x1, x2, x3; /* temporary variables for holding input and coefficient values */
emh203 0:3d9c67d97d6f 97 uint32_t j, blkCnt, outBlockSize; /* loop counter */
emh203 0:3d9c67d97d6f 98 int32_t inc = 1; /* Destination address modifier */
emh203 0:3d9c67d97d6f 99 uint32_t tapCnt;
emh203 0:3d9c67d97d6f 100 q31_t y1, y2;
emh203 0:3d9c67d97d6f 101 q15_t *pScr; /* Intermediate pointers */
emh203 0:3d9c67d97d6f 102 q15_t *pOut = pDst; /* output pointer */
emh203 0:3d9c67d97d6f 103 #ifdef UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 104
emh203 0:3d9c67d97d6f 105 q15_t a, b;
emh203 0:3d9c67d97d6f 106
emh203 0:3d9c67d97d6f 107 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emh203 0:3d9c67d97d6f 108
emh203 0:3d9c67d97d6f 109 /* The algorithm implementation is based on the lengths of the inputs. */
emh203 0:3d9c67d97d6f 110 /* srcB is always made to slide across srcA. */
emh203 0:3d9c67d97d6f 111 /* So srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 112 /* But CORR(x, y) is reverse of CORR(y, x) */
emh203 0:3d9c67d97d6f 113 /* So, when srcBLen > srcALen, output pointer is made to point to the end of the output buffer */
emh203 0:3d9c67d97d6f 114 /* and the destination pointer modifier, inc is set to -1 */
emh203 0:3d9c67d97d6f 115 /* If srcALen > srcBLen, zero pad has to be done to srcB to make the two inputs of same length */
emh203 0:3d9c67d97d6f 116 /* But to improve the performance,
emh203 0:3d9c67d97d6f 117 * we include zeroes in the output instead of zero padding either of the the inputs*/
emh203 0:3d9c67d97d6f 118 /* If srcALen > srcBLen,
emh203 0:3d9c67d97d6f 119 * (srcALen - srcBLen) zeroes has to included in the starting of the output buffer */
emh203 0:3d9c67d97d6f 120 /* If srcALen < srcBLen,
emh203 0:3d9c67d97d6f 121 * (srcALen - srcBLen) zeroes has to included in the ending of the output buffer */
emh203 0:3d9c67d97d6f 122 if(srcALen >= srcBLen)
emh203 0:3d9c67d97d6f 123 {
emh203 0:3d9c67d97d6f 124 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 125 pIn1 = (pSrcA);
emh203 0:3d9c67d97d6f 126
emh203 0:3d9c67d97d6f 127 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 128 pIn2 = (pSrcB);
emh203 0:3d9c67d97d6f 129
emh203 0:3d9c67d97d6f 130 /* Number of output samples is calculated */
emh203 0:3d9c67d97d6f 131 outBlockSize = (2u * srcALen) - 1u;
emh203 0:3d9c67d97d6f 132
emh203 0:3d9c67d97d6f 133 /* When srcALen > srcBLen, zero padding is done to srcB
emh203 0:3d9c67d97d6f 134 * to make their lengths equal.
emh203 0:3d9c67d97d6f 135 * Instead, (outBlockSize - (srcALen + srcBLen - 1))
emh203 0:3d9c67d97d6f 136 * number of output samples are made zero */
emh203 0:3d9c67d97d6f 137 j = outBlockSize - (srcALen + (srcBLen - 1u));
emh203 0:3d9c67d97d6f 138
emh203 0:3d9c67d97d6f 139 /* Updating the pointer position to non zero value */
emh203 0:3d9c67d97d6f 140 pOut += j;
emh203 0:3d9c67d97d6f 141
emh203 0:3d9c67d97d6f 142 }
emh203 0:3d9c67d97d6f 143 else
emh203 0:3d9c67d97d6f 144 {
emh203 0:3d9c67d97d6f 145 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 146 pIn1 = (pSrcB);
emh203 0:3d9c67d97d6f 147
emh203 0:3d9c67d97d6f 148 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 149 pIn2 = (pSrcA);
emh203 0:3d9c67d97d6f 150
emh203 0:3d9c67d97d6f 151 /* srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 152 j = srcBLen;
emh203 0:3d9c67d97d6f 153 srcBLen = srcALen;
emh203 0:3d9c67d97d6f 154 srcALen = j;
emh203 0:3d9c67d97d6f 155
emh203 0:3d9c67d97d6f 156 /* CORR(x, y) = Reverse order(CORR(y, x)) */
emh203 0:3d9c67d97d6f 157 /* Hence set the destination pointer to point to the last output sample */
emh203 0:3d9c67d97d6f 158 pOut = pDst + ((srcALen + srcBLen) - 2u);
emh203 0:3d9c67d97d6f 159
emh203 0:3d9c67d97d6f 160 /* Destination address modifier is set to -1 */
emh203 0:3d9c67d97d6f 161 inc = -1;
emh203 0:3d9c67d97d6f 162
emh203 0:3d9c67d97d6f 163 }
emh203 0:3d9c67d97d6f 164
emh203 0:3d9c67d97d6f 165 pScr = pScratch;
emh203 0:3d9c67d97d6f 166
emh203 0:3d9c67d97d6f 167 /* Fill (srcBLen - 1u) zeros in scratch buffer */
emh203 0:3d9c67d97d6f 168 arm_fill_q15(0, pScr, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 169
emh203 0:3d9c67d97d6f 170 /* Update temporary scratch pointer */
emh203 0:3d9c67d97d6f 171 pScr += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 172
emh203 0:3d9c67d97d6f 173 #ifndef UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 174
emh203 0:3d9c67d97d6f 175 /* Copy (srcALen) samples in scratch buffer */
emh203 0:3d9c67d97d6f 176 arm_copy_q15(pIn1, pScr, srcALen);
emh203 0:3d9c67d97d6f 177
emh203 0:3d9c67d97d6f 178 /* Update pointers */
emh203 0:3d9c67d97d6f 179 pScr += srcALen;
emh203 0:3d9c67d97d6f 180
emh203 0:3d9c67d97d6f 181 #else
emh203 0:3d9c67d97d6f 182
emh203 0:3d9c67d97d6f 183 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 184 j = srcALen >> 2u;
emh203 0:3d9c67d97d6f 185
emh203 0:3d9c67d97d6f 186 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 187 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 188 while(j > 0u)
emh203 0:3d9c67d97d6f 189 {
emh203 0:3d9c67d97d6f 190 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 191 *pScr++ = *pIn1++;
emh203 0:3d9c67d97d6f 192 *pScr++ = *pIn1++;
emh203 0:3d9c67d97d6f 193 *pScr++ = *pIn1++;
emh203 0:3d9c67d97d6f 194 *pScr++ = *pIn1++;
emh203 0:3d9c67d97d6f 195
emh203 0:3d9c67d97d6f 196 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 197 j--;
emh203 0:3d9c67d97d6f 198 }
emh203 0:3d9c67d97d6f 199
emh203 0:3d9c67d97d6f 200 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 201 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 202 j = srcALen % 0x4u;
emh203 0:3d9c67d97d6f 203
emh203 0:3d9c67d97d6f 204 while(j > 0u)
emh203 0:3d9c67d97d6f 205 {
emh203 0:3d9c67d97d6f 206 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 207 *pScr++ = *pIn1++;
emh203 0:3d9c67d97d6f 208
emh203 0:3d9c67d97d6f 209 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 210 j--;
emh203 0:3d9c67d97d6f 211 }
emh203 0:3d9c67d97d6f 212
emh203 0:3d9c67d97d6f 213 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emh203 0:3d9c67d97d6f 214
emh203 0:3d9c67d97d6f 215 #ifndef UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 216
emh203 0:3d9c67d97d6f 217 /* Fill (srcBLen - 1u) zeros at end of scratch buffer */
emh203 0:3d9c67d97d6f 218 arm_fill_q15(0, pScr, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 219
emh203 0:3d9c67d97d6f 220 /* Update pointer */
emh203 0:3d9c67d97d6f 221 pScr += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 222
emh203 0:3d9c67d97d6f 223 #else
emh203 0:3d9c67d97d6f 224
emh203 0:3d9c67d97d6f 225 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 226 j = (srcBLen - 1u) >> 2u;
emh203 0:3d9c67d97d6f 227
emh203 0:3d9c67d97d6f 228 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 229 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 230 while(j > 0u)
emh203 0:3d9c67d97d6f 231 {
emh203 0:3d9c67d97d6f 232 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 233 *pScr++ = 0;
emh203 0:3d9c67d97d6f 234 *pScr++ = 0;
emh203 0:3d9c67d97d6f 235 *pScr++ = 0;
emh203 0:3d9c67d97d6f 236 *pScr++ = 0;
emh203 0:3d9c67d97d6f 237
emh203 0:3d9c67d97d6f 238 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 239 j--;
emh203 0:3d9c67d97d6f 240 }
emh203 0:3d9c67d97d6f 241
emh203 0:3d9c67d97d6f 242 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 243 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 244 j = (srcBLen - 1u) % 0x4u;
emh203 0:3d9c67d97d6f 245
emh203 0:3d9c67d97d6f 246 while(j > 0u)
emh203 0:3d9c67d97d6f 247 {
emh203 0:3d9c67d97d6f 248 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 249 *pScr++ = 0;
emh203 0:3d9c67d97d6f 250
emh203 0:3d9c67d97d6f 251 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 252 j--;
emh203 0:3d9c67d97d6f 253 }
emh203 0:3d9c67d97d6f 254
emh203 0:3d9c67d97d6f 255 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emh203 0:3d9c67d97d6f 256
emh203 0:3d9c67d97d6f 257 /* Temporary pointer for scratch2 */
emh203 0:3d9c67d97d6f 258 py = pIn2;
emh203 0:3d9c67d97d6f 259
emh203 0:3d9c67d97d6f 260
emh203 0:3d9c67d97d6f 261 /* Actual correlation process starts here */
emh203 0:3d9c67d97d6f 262 blkCnt = (srcALen + srcBLen - 1u) >> 2;
emh203 0:3d9c67d97d6f 263
emh203 0:3d9c67d97d6f 264 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 265 {
emh203 0:3d9c67d97d6f 266 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 267 pScr = pScratch;
emh203 0:3d9c67d97d6f 268
emh203 0:3d9c67d97d6f 269 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 270 acc0 = 0;
emh203 0:3d9c67d97d6f 271 acc1 = 0;
emh203 0:3d9c67d97d6f 272 acc2 = 0;
emh203 0:3d9c67d97d6f 273 acc3 = 0;
emh203 0:3d9c67d97d6f 274
emh203 0:3d9c67d97d6f 275 /* Read four samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 276 x1 = *__SIMD32(pScr)++;
emh203 0:3d9c67d97d6f 277
emh203 0:3d9c67d97d6f 278 /* Read next four samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 279 x2 = *__SIMD32(pScr)++;
emh203 0:3d9c67d97d6f 280
emh203 0:3d9c67d97d6f 281 tapCnt = (srcBLen) >> 2u;
emh203 0:3d9c67d97d6f 282
emh203 0:3d9c67d97d6f 283 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 284 {
emh203 0:3d9c67d97d6f 285
emh203 0:3d9c67d97d6f 286 #ifndef UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 287
emh203 0:3d9c67d97d6f 288 /* Read four samples from smaller buffer */
emh203 0:3d9c67d97d6f 289 y1 = _SIMD32_OFFSET(pIn2);
emh203 0:3d9c67d97d6f 290 y2 = _SIMD32_OFFSET(pIn2 + 2u);
emh203 0:3d9c67d97d6f 291
emh203 0:3d9c67d97d6f 292 acc0 = __SMLAD(x1, y1, acc0);
emh203 0:3d9c67d97d6f 293
emh203 0:3d9c67d97d6f 294 acc2 = __SMLAD(x2, y1, acc2);
emh203 0:3d9c67d97d6f 295
emh203 0:3d9c67d97d6f 296 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 297 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 298 #else
emh203 0:3d9c67d97d6f 299 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 300 #endif
emh203 0:3d9c67d97d6f 301
emh203 0:3d9c67d97d6f 302 acc1 = __SMLADX(x3, y1, acc1);
emh203 0:3d9c67d97d6f 303
emh203 0:3d9c67d97d6f 304 x1 = _SIMD32_OFFSET(pScr);
emh203 0:3d9c67d97d6f 305
emh203 0:3d9c67d97d6f 306 acc0 = __SMLAD(x2, y2, acc0);
emh203 0:3d9c67d97d6f 307
emh203 0:3d9c67d97d6f 308 acc2 = __SMLAD(x1, y2, acc2);
emh203 0:3d9c67d97d6f 309
emh203 0:3d9c67d97d6f 310 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 311 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 312 #else
emh203 0:3d9c67d97d6f 313 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 314 #endif
emh203 0:3d9c67d97d6f 315
emh203 0:3d9c67d97d6f 316 acc3 = __SMLADX(x3, y1, acc3);
emh203 0:3d9c67d97d6f 317
emh203 0:3d9c67d97d6f 318 acc1 = __SMLADX(x3, y2, acc1);
emh203 0:3d9c67d97d6f 319
emh203 0:3d9c67d97d6f 320 x2 = _SIMD32_OFFSET(pScr + 2u);
emh203 0:3d9c67d97d6f 321
emh203 0:3d9c67d97d6f 322 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 323 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 324 #else
emh203 0:3d9c67d97d6f 325 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 326 #endif
emh203 0:3d9c67d97d6f 327
emh203 0:3d9c67d97d6f 328 acc3 = __SMLADX(x3, y2, acc3);
emh203 0:3d9c67d97d6f 329 #else
emh203 0:3d9c67d97d6f 330
emh203 0:3d9c67d97d6f 331 /* Read four samples from smaller buffer */
emh203 0:3d9c67d97d6f 332 a = *pIn2;
emh203 0:3d9c67d97d6f 333 b = *(pIn2 + 1);
emh203 0:3d9c67d97d6f 334
emh203 0:3d9c67d97d6f 335 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 336 y1 = __PKHBT(a, b, 16);
emh203 0:3d9c67d97d6f 337 #else
emh203 0:3d9c67d97d6f 338 y1 = __PKHBT(b, a, 16);
emh203 0:3d9c67d97d6f 339 #endif
emh203 0:3d9c67d97d6f 340
emh203 0:3d9c67d97d6f 341 a = *(pIn2 + 2);
emh203 0:3d9c67d97d6f 342 b = *(pIn2 + 3);
emh203 0:3d9c67d97d6f 343 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 344 y2 = __PKHBT(a, b, 16);
emh203 0:3d9c67d97d6f 345 #else
emh203 0:3d9c67d97d6f 346 y2 = __PKHBT(b, a, 16);
emh203 0:3d9c67d97d6f 347 #endif
emh203 0:3d9c67d97d6f 348
emh203 0:3d9c67d97d6f 349 acc0 = __SMLAD(x1, y1, acc0);
emh203 0:3d9c67d97d6f 350
emh203 0:3d9c67d97d6f 351 acc2 = __SMLAD(x2, y1, acc2);
emh203 0:3d9c67d97d6f 352
emh203 0:3d9c67d97d6f 353 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 354 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 355 #else
emh203 0:3d9c67d97d6f 356 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 357 #endif
emh203 0:3d9c67d97d6f 358
emh203 0:3d9c67d97d6f 359 acc1 = __SMLADX(x3, y1, acc1);
emh203 0:3d9c67d97d6f 360
emh203 0:3d9c67d97d6f 361 a = *pScr;
emh203 0:3d9c67d97d6f 362 b = *(pScr + 1);
emh203 0:3d9c67d97d6f 363
emh203 0:3d9c67d97d6f 364 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 365 x1 = __PKHBT(a, b, 16);
emh203 0:3d9c67d97d6f 366 #else
emh203 0:3d9c67d97d6f 367 x1 = __PKHBT(b, a, 16);
emh203 0:3d9c67d97d6f 368 #endif
emh203 0:3d9c67d97d6f 369
emh203 0:3d9c67d97d6f 370 acc0 = __SMLAD(x2, y2, acc0);
emh203 0:3d9c67d97d6f 371
emh203 0:3d9c67d97d6f 372 acc2 = __SMLAD(x1, y2, acc2);
emh203 0:3d9c67d97d6f 373
emh203 0:3d9c67d97d6f 374 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 375 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 376 #else
emh203 0:3d9c67d97d6f 377 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 378 #endif
emh203 0:3d9c67d97d6f 379
emh203 0:3d9c67d97d6f 380 acc3 = __SMLADX(x3, y1, acc3);
emh203 0:3d9c67d97d6f 381
emh203 0:3d9c67d97d6f 382 acc1 = __SMLADX(x3, y2, acc1);
emh203 0:3d9c67d97d6f 383
emh203 0:3d9c67d97d6f 384 a = *(pScr + 2);
emh203 0:3d9c67d97d6f 385 b = *(pScr + 3);
emh203 0:3d9c67d97d6f 386
emh203 0:3d9c67d97d6f 387 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 388 x2 = __PKHBT(a, b, 16);
emh203 0:3d9c67d97d6f 389 #else
emh203 0:3d9c67d97d6f 390 x2 = __PKHBT(b, a, 16);
emh203 0:3d9c67d97d6f 391 #endif
emh203 0:3d9c67d97d6f 392
emh203 0:3d9c67d97d6f 393 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 394 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 395 #else
emh203 0:3d9c67d97d6f 396 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 397 #endif
emh203 0:3d9c67d97d6f 398
emh203 0:3d9c67d97d6f 399 acc3 = __SMLADX(x3, y2, acc3);
emh203 0:3d9c67d97d6f 400
emh203 0:3d9c67d97d6f 401 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emh203 0:3d9c67d97d6f 402
emh203 0:3d9c67d97d6f 403 pIn2 += 4u;
emh203 0:3d9c67d97d6f 404
emh203 0:3d9c67d97d6f 405 pScr += 4u;
emh203 0:3d9c67d97d6f 406
emh203 0:3d9c67d97d6f 407
emh203 0:3d9c67d97d6f 408 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 409 tapCnt--;
emh203 0:3d9c67d97d6f 410 }
emh203 0:3d9c67d97d6f 411
emh203 0:3d9c67d97d6f 412
emh203 0:3d9c67d97d6f 413
emh203 0:3d9c67d97d6f 414 /* Update scratch pointer for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 415 pScr -= 4u;
emh203 0:3d9c67d97d6f 416
emh203 0:3d9c67d97d6f 417
emh203 0:3d9c67d97d6f 418 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 419 tapCnt = (srcBLen) & 3u;
emh203 0:3d9c67d97d6f 420
emh203 0:3d9c67d97d6f 421 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 422 {
emh203 0:3d9c67d97d6f 423
emh203 0:3d9c67d97d6f 424 /* accumlate the results */
emh203 0:3d9c67d97d6f 425 acc0 += (*pScr++ * *pIn2);
emh203 0:3d9c67d97d6f 426 acc1 += (*pScr++ * *pIn2);
emh203 0:3d9c67d97d6f 427 acc2 += (*pScr++ * *pIn2);
emh203 0:3d9c67d97d6f 428 acc3 += (*pScr++ * *pIn2++);
emh203 0:3d9c67d97d6f 429
emh203 0:3d9c67d97d6f 430 pScr -= 3u;
emh203 0:3d9c67d97d6f 431
emh203 0:3d9c67d97d6f 432 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 433 tapCnt--;
emh203 0:3d9c67d97d6f 434 }
emh203 0:3d9c67d97d6f 435
emh203 0:3d9c67d97d6f 436 blkCnt--;
emh203 0:3d9c67d97d6f 437
emh203 0:3d9c67d97d6f 438
emh203 0:3d9c67d97d6f 439 /* Store the results in the accumulators in the destination buffer. */
emh203 0:3d9c67d97d6f 440 *pOut = (__SSAT(acc0 >> 15u, 16));
emh203 0:3d9c67d97d6f 441 pOut += inc;
emh203 0:3d9c67d97d6f 442 *pOut = (__SSAT(acc1 >> 15u, 16));
emh203 0:3d9c67d97d6f 443 pOut += inc;
emh203 0:3d9c67d97d6f 444 *pOut = (__SSAT(acc2 >> 15u, 16));
emh203 0:3d9c67d97d6f 445 pOut += inc;
emh203 0:3d9c67d97d6f 446 *pOut = (__SSAT(acc3 >> 15u, 16));
emh203 0:3d9c67d97d6f 447 pOut += inc;
emh203 0:3d9c67d97d6f 448
emh203 0:3d9c67d97d6f 449
emh203 0:3d9c67d97d6f 450 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 451 pIn2 = py;
emh203 0:3d9c67d97d6f 452
emh203 0:3d9c67d97d6f 453 pScratch += 4u;
emh203 0:3d9c67d97d6f 454
emh203 0:3d9c67d97d6f 455 }
emh203 0:3d9c67d97d6f 456
emh203 0:3d9c67d97d6f 457
emh203 0:3d9c67d97d6f 458 blkCnt = (srcALen + srcBLen - 1u) & 0x3;
emh203 0:3d9c67d97d6f 459
emh203 0:3d9c67d97d6f 460 /* Calculate correlation for remaining samples of Bigger length sequence */
emh203 0:3d9c67d97d6f 461 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 462 {
emh203 0:3d9c67d97d6f 463 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 464 pScr = pScratch;
emh203 0:3d9c67d97d6f 465
emh203 0:3d9c67d97d6f 466 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 467 acc0 = 0;
emh203 0:3d9c67d97d6f 468
emh203 0:3d9c67d97d6f 469 tapCnt = (srcBLen) >> 1u;
emh203 0:3d9c67d97d6f 470
emh203 0:3d9c67d97d6f 471 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 472 {
emh203 0:3d9c67d97d6f 473
emh203 0:3d9c67d97d6f 474 acc0 += (*pScr++ * *pIn2++);
emh203 0:3d9c67d97d6f 475 acc0 += (*pScr++ * *pIn2++);
emh203 0:3d9c67d97d6f 476
emh203 0:3d9c67d97d6f 477 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 478 tapCnt--;
emh203 0:3d9c67d97d6f 479 }
emh203 0:3d9c67d97d6f 480
emh203 0:3d9c67d97d6f 481 tapCnt = (srcBLen) & 1u;
emh203 0:3d9c67d97d6f 482
emh203 0:3d9c67d97d6f 483 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 484 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 485 {
emh203 0:3d9c67d97d6f 486
emh203 0:3d9c67d97d6f 487 /* accumlate the results */
emh203 0:3d9c67d97d6f 488 acc0 += (*pScr++ * *pIn2++);
emh203 0:3d9c67d97d6f 489
emh203 0:3d9c67d97d6f 490 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 491 tapCnt--;
emh203 0:3d9c67d97d6f 492 }
emh203 0:3d9c67d97d6f 493
emh203 0:3d9c67d97d6f 494 blkCnt--;
emh203 0:3d9c67d97d6f 495
emh203 0:3d9c67d97d6f 496 /* Store the result in the accumulator in the destination buffer. */
emh203 0:3d9c67d97d6f 497
emh203 0:3d9c67d97d6f 498 *pOut = (q15_t) (__SSAT((acc0 >> 15), 16));
emh203 0:3d9c67d97d6f 499
emh203 0:3d9c67d97d6f 500 pOut += inc;
emh203 0:3d9c67d97d6f 501
emh203 0:3d9c67d97d6f 502 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 503 pIn2 = py;
emh203 0:3d9c67d97d6f 504
emh203 0:3d9c67d97d6f 505 pScratch += 1u;
emh203 0:3d9c67d97d6f 506
emh203 0:3d9c67d97d6f 507 }
emh203 0:3d9c67d97d6f 508 }
emh203 0:3d9c67d97d6f 509
emh203 0:3d9c67d97d6f 510 /**
emh203 0:3d9c67d97d6f 511 * @} end of Corr group
emh203 0:3d9c67d97d6f 512 */