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_opt_q7.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_opt_q7.c
emh203 0:3d9c67d97d6f 9 *
emh203 0:3d9c67d97d6f 10 * Description: Correlation of Q7 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 Corr
emh203 0:3d9c67d97d6f 49 * @{
emh203 0:3d9c67d97d6f 50 */
emh203 0:3d9c67d97d6f 51
emh203 0:3d9c67d97d6f 52 /**
emh203 0:3d9c67d97d6f 53 * @brief Correlation of Q7 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. Length 2 * max(srcALen, srcBLen) - 1.
emh203 0:3d9c67d97d6f 59 * @param[in] *pScratch1 points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
emh203 0:3d9c67d97d6f 60 * @param[in] *pScratch2 points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
emh203 0:3d9c67d97d6f 61 * @return none.
emh203 0:3d9c67d97d6f 62 *
emh203 0:3d9c67d97d6f 63 *
emh203 0:3d9c67d97d6f 64 * \par Restrictions
emh203 0:3d9c67d97d6f 65 * If the silicon does not support unaligned memory access enable the macro UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 66 * In this case input, output, scratch1 and scratch2 buffers should be aligned by 32-bit
emh203 0:3d9c67d97d6f 67 *
emh203 0:3d9c67d97d6f 68 * @details
emh203 0:3d9c67d97d6f 69 * <b>Scaling and Overflow Behavior:</b>
emh203 0:3d9c67d97d6f 70 *
emh203 0:3d9c67d97d6f 71 * \par
emh203 0:3d9c67d97d6f 72 * The function is implemented using a 32-bit internal accumulator.
emh203 0:3d9c67d97d6f 73 * Both the inputs are represented in 1.7 format and multiplications yield a 2.14 result.
emh203 0:3d9c67d97d6f 74 * The 2.14 intermediate results are accumulated in a 32-bit accumulator in 18.14 format.
emh203 0:3d9c67d97d6f 75 * This approach provides 17 guard bits and there is no risk of overflow as long as <code>max(srcALen, srcBLen)<131072</code>.
emh203 0:3d9c67d97d6f 76 * The 18.14 result is then truncated to 18.7 format by discarding the low 7 bits and saturated to 1.7 format.
emh203 0:3d9c67d97d6f 77 *
emh203 0:3d9c67d97d6f 78 *
emh203 0:3d9c67d97d6f 79 */
emh203 0:3d9c67d97d6f 80
emh203 0:3d9c67d97d6f 81
emh203 0:3d9c67d97d6f 82
emh203 0:3d9c67d97d6f 83 void arm_correlate_opt_q7(
emh203 0:3d9c67d97d6f 84 q7_t * pSrcA,
emh203 0:3d9c67d97d6f 85 uint32_t srcALen,
emh203 0:3d9c67d97d6f 86 q7_t * pSrcB,
emh203 0:3d9c67d97d6f 87 uint32_t srcBLen,
emh203 0:3d9c67d97d6f 88 q7_t * pDst,
emh203 0:3d9c67d97d6f 89 q15_t * pScratch1,
emh203 0:3d9c67d97d6f 90 q15_t * pScratch2)
emh203 0:3d9c67d97d6f 91 {
emh203 0:3d9c67d97d6f 92 q7_t *pOut = pDst; /* output pointer */
emh203 0:3d9c67d97d6f 93 q15_t *pScr1 = pScratch1; /* Temporary pointer for scratch */
emh203 0:3d9c67d97d6f 94 q15_t *pScr2 = pScratch2; /* Temporary pointer for scratch */
emh203 0:3d9c67d97d6f 95 q7_t *pIn1; /* inputA pointer */
emh203 0:3d9c67d97d6f 96 q7_t *pIn2; /* inputB pointer */
emh203 0:3d9c67d97d6f 97 q15_t *py; /* Intermediate inputB pointer */
emh203 0:3d9c67d97d6f 98 q31_t acc0, acc1, acc2, acc3; /* Accumulators */
emh203 0:3d9c67d97d6f 99 uint32_t j, k = 0u, blkCnt; /* loop counter */
emh203 0:3d9c67d97d6f 100 int32_t inc = 1; /* output pointer increment */
emh203 0:3d9c67d97d6f 101 uint32_t outBlockSize; /* loop counter */
emh203 0:3d9c67d97d6f 102 q15_t x4; /* Temporary input variable */
emh203 0:3d9c67d97d6f 103 uint32_t tapCnt; /* loop counter */
emh203 0:3d9c67d97d6f 104 q31_t x1, x2, x3, y1; /* Temporary input variables */
emh203 0:3d9c67d97d6f 105
emh203 0:3d9c67d97d6f 106 /* The algorithm implementation is based on the lengths of the inputs. */
emh203 0:3d9c67d97d6f 107 /* srcB is always made to slide across srcA. */
emh203 0:3d9c67d97d6f 108 /* So srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 109 /* But CORR(x, y) is reverse of CORR(y, x) */
emh203 0:3d9c67d97d6f 110 /* So, when srcBLen > srcALen, output pointer is made to point to the end of the output buffer */
emh203 0:3d9c67d97d6f 111 /* and the destination pointer modifier, inc is set to -1 */
emh203 0:3d9c67d97d6f 112 /* If srcALen > srcBLen, zero pad has to be done to srcB to make the two inputs of same length */
emh203 0:3d9c67d97d6f 113 /* But to improve the performance,
emh203 0:3d9c67d97d6f 114 * we include zeroes in the output instead of zero padding either of the the inputs*/
emh203 0:3d9c67d97d6f 115 /* If srcALen > srcBLen,
emh203 0:3d9c67d97d6f 116 * (srcALen - srcBLen) zeroes has to included in the starting of the output buffer */
emh203 0:3d9c67d97d6f 117 /* If srcALen < srcBLen,
emh203 0:3d9c67d97d6f 118 * (srcALen - srcBLen) zeroes has to included in the ending of the output buffer */
emh203 0:3d9c67d97d6f 119 if(srcALen >= srcBLen)
emh203 0:3d9c67d97d6f 120 {
emh203 0:3d9c67d97d6f 121 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 122 pIn1 = (pSrcA);
emh203 0:3d9c67d97d6f 123
emh203 0:3d9c67d97d6f 124 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 125 pIn2 = (pSrcB);
emh203 0:3d9c67d97d6f 126
emh203 0:3d9c67d97d6f 127 /* Number of output samples is calculated */
emh203 0:3d9c67d97d6f 128 outBlockSize = (2u * srcALen) - 1u;
emh203 0:3d9c67d97d6f 129
emh203 0:3d9c67d97d6f 130 /* When srcALen > srcBLen, zero padding is done to srcB
emh203 0:3d9c67d97d6f 131 * to make their lengths equal.
emh203 0:3d9c67d97d6f 132 * Instead, (outBlockSize - (srcALen + srcBLen - 1))
emh203 0:3d9c67d97d6f 133 * number of output samples are made zero */
emh203 0:3d9c67d97d6f 134 j = outBlockSize - (srcALen + (srcBLen - 1u));
emh203 0:3d9c67d97d6f 135
emh203 0:3d9c67d97d6f 136 /* Updating the pointer position to non zero value */
emh203 0:3d9c67d97d6f 137 pOut += j;
emh203 0:3d9c67d97d6f 138
emh203 0:3d9c67d97d6f 139 }
emh203 0:3d9c67d97d6f 140 else
emh203 0:3d9c67d97d6f 141 {
emh203 0:3d9c67d97d6f 142 /* Initialization of inputA pointer */
emh203 0:3d9c67d97d6f 143 pIn1 = (pSrcB);
emh203 0:3d9c67d97d6f 144
emh203 0:3d9c67d97d6f 145 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 146 pIn2 = (pSrcA);
emh203 0:3d9c67d97d6f 147
emh203 0:3d9c67d97d6f 148 /* srcBLen is always considered as shorter or equal to srcALen */
emh203 0:3d9c67d97d6f 149 j = srcBLen;
emh203 0:3d9c67d97d6f 150 srcBLen = srcALen;
emh203 0:3d9c67d97d6f 151 srcALen = j;
emh203 0:3d9c67d97d6f 152
emh203 0:3d9c67d97d6f 153 /* CORR(x, y) = Reverse order(CORR(y, x)) */
emh203 0:3d9c67d97d6f 154 /* Hence set the destination pointer to point to the last output sample */
emh203 0:3d9c67d97d6f 155 pOut = pDst + ((srcALen + srcBLen) - 2u);
emh203 0:3d9c67d97d6f 156
emh203 0:3d9c67d97d6f 157 /* Destination address modifier is set to -1 */
emh203 0:3d9c67d97d6f 158 inc = -1;
emh203 0:3d9c67d97d6f 159
emh203 0:3d9c67d97d6f 160 }
emh203 0:3d9c67d97d6f 161
emh203 0:3d9c67d97d6f 162
emh203 0:3d9c67d97d6f 163 /* Copy (srcBLen) samples in scratch buffer */
emh203 0:3d9c67d97d6f 164 k = srcBLen >> 2u;
emh203 0:3d9c67d97d6f 165
emh203 0:3d9c67d97d6f 166 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 167 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 168 while(k > 0u)
emh203 0:3d9c67d97d6f 169 {
emh203 0:3d9c67d97d6f 170 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 171 x4 = (q15_t) * pIn2++;
emh203 0:3d9c67d97d6f 172 *pScr2++ = x4;
emh203 0:3d9c67d97d6f 173 x4 = (q15_t) * pIn2++;
emh203 0:3d9c67d97d6f 174 *pScr2++ = x4;
emh203 0:3d9c67d97d6f 175 x4 = (q15_t) * pIn2++;
emh203 0:3d9c67d97d6f 176 *pScr2++ = x4;
emh203 0:3d9c67d97d6f 177 x4 = (q15_t) * pIn2++;
emh203 0:3d9c67d97d6f 178 *pScr2++ = x4;
emh203 0:3d9c67d97d6f 179
emh203 0:3d9c67d97d6f 180 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 181 k--;
emh203 0:3d9c67d97d6f 182 }
emh203 0:3d9c67d97d6f 183
emh203 0:3d9c67d97d6f 184 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 185 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 186 k = srcBLen % 0x4u;
emh203 0:3d9c67d97d6f 187
emh203 0:3d9c67d97d6f 188 while(k > 0u)
emh203 0:3d9c67d97d6f 189 {
emh203 0:3d9c67d97d6f 190 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 191 x4 = (q15_t) * pIn2++;
emh203 0:3d9c67d97d6f 192 *pScr2++ = x4;
emh203 0:3d9c67d97d6f 193
emh203 0:3d9c67d97d6f 194 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 195 k--;
emh203 0:3d9c67d97d6f 196 }
emh203 0:3d9c67d97d6f 197
emh203 0:3d9c67d97d6f 198 /* Fill (srcBLen - 1u) zeros in scratch buffer */
emh203 0:3d9c67d97d6f 199 arm_fill_q15(0, pScr1, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 200
emh203 0:3d9c67d97d6f 201 /* Update temporary scratch pointer */
emh203 0:3d9c67d97d6f 202 pScr1 += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 203
emh203 0:3d9c67d97d6f 204 /* Copy (srcALen) samples in scratch buffer */
emh203 0:3d9c67d97d6f 205 k = srcALen >> 2u;
emh203 0:3d9c67d97d6f 206
emh203 0:3d9c67d97d6f 207 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 208 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 209 while(k > 0u)
emh203 0:3d9c67d97d6f 210 {
emh203 0:3d9c67d97d6f 211 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 212 x4 = (q15_t) * pIn1++;
emh203 0:3d9c67d97d6f 213 *pScr1++ = x4;
emh203 0:3d9c67d97d6f 214 x4 = (q15_t) * pIn1++;
emh203 0:3d9c67d97d6f 215 *pScr1++ = x4;
emh203 0:3d9c67d97d6f 216 x4 = (q15_t) * pIn1++;
emh203 0:3d9c67d97d6f 217 *pScr1++ = x4;
emh203 0:3d9c67d97d6f 218 x4 = (q15_t) * pIn1++;
emh203 0:3d9c67d97d6f 219 *pScr1++ = x4;
emh203 0:3d9c67d97d6f 220
emh203 0:3d9c67d97d6f 221 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 222 k--;
emh203 0:3d9c67d97d6f 223 }
emh203 0:3d9c67d97d6f 224
emh203 0:3d9c67d97d6f 225 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 226 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 227 k = srcALen % 0x4u;
emh203 0:3d9c67d97d6f 228
emh203 0:3d9c67d97d6f 229 while(k > 0u)
emh203 0:3d9c67d97d6f 230 {
emh203 0:3d9c67d97d6f 231 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 232 x4 = (q15_t) * pIn1++;
emh203 0:3d9c67d97d6f 233 *pScr1++ = x4;
emh203 0:3d9c67d97d6f 234
emh203 0:3d9c67d97d6f 235 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 236 k--;
emh203 0:3d9c67d97d6f 237 }
emh203 0:3d9c67d97d6f 238
emh203 0:3d9c67d97d6f 239 #ifndef UNALIGNED_SUPPORT_DISABLE
emh203 0:3d9c67d97d6f 240
emh203 0:3d9c67d97d6f 241 /* Fill (srcBLen - 1u) zeros at end of scratch buffer */
emh203 0:3d9c67d97d6f 242 arm_fill_q15(0, pScr1, (srcBLen - 1u));
emh203 0:3d9c67d97d6f 243
emh203 0:3d9c67d97d6f 244 /* Update pointer */
emh203 0:3d9c67d97d6f 245 pScr1 += (srcBLen - 1u);
emh203 0:3d9c67d97d6f 246
emh203 0:3d9c67d97d6f 247 #else
emh203 0:3d9c67d97d6f 248
emh203 0:3d9c67d97d6f 249 /* Apply loop unrolling and do 4 Copies simultaneously. */
emh203 0:3d9c67d97d6f 250 k = (srcBLen - 1u) >> 2u;
emh203 0:3d9c67d97d6f 251
emh203 0:3d9c67d97d6f 252 /* First part of the processing with loop unrolling copies 4 data points at a time.
emh203 0:3d9c67d97d6f 253 ** a second loop below copies for the remaining 1 to 3 samples. */
emh203 0:3d9c67d97d6f 254 while(k > 0u)
emh203 0:3d9c67d97d6f 255 {
emh203 0:3d9c67d97d6f 256 /* copy second buffer in reversal manner */
emh203 0:3d9c67d97d6f 257 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 258 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 259 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 260 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 261
emh203 0:3d9c67d97d6f 262 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 263 k--;
emh203 0:3d9c67d97d6f 264 }
emh203 0:3d9c67d97d6f 265
emh203 0:3d9c67d97d6f 266 /* If the count is not a multiple of 4, copy remaining samples here.
emh203 0:3d9c67d97d6f 267 ** No loop unrolling is used. */
emh203 0:3d9c67d97d6f 268 k = (srcBLen - 1u) % 0x4u;
emh203 0:3d9c67d97d6f 269
emh203 0:3d9c67d97d6f 270 while(k > 0u)
emh203 0:3d9c67d97d6f 271 {
emh203 0:3d9c67d97d6f 272 /* copy second buffer in reversal manner for remaining samples */
emh203 0:3d9c67d97d6f 273 *pScr1++ = 0;
emh203 0:3d9c67d97d6f 274
emh203 0:3d9c67d97d6f 275 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 276 k--;
emh203 0:3d9c67d97d6f 277 }
emh203 0:3d9c67d97d6f 278
emh203 0:3d9c67d97d6f 279 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emh203 0:3d9c67d97d6f 280
emh203 0:3d9c67d97d6f 281 /* Temporary pointer for second sequence */
emh203 0:3d9c67d97d6f 282 py = pScratch2;
emh203 0:3d9c67d97d6f 283
emh203 0:3d9c67d97d6f 284 /* Initialization of pScr2 pointer */
emh203 0:3d9c67d97d6f 285 pScr2 = pScratch2;
emh203 0:3d9c67d97d6f 286
emh203 0:3d9c67d97d6f 287 /* Actual correlation process starts here */
emh203 0:3d9c67d97d6f 288 blkCnt = (srcALen + srcBLen - 1u) >> 2;
emh203 0:3d9c67d97d6f 289
emh203 0:3d9c67d97d6f 290 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 291 {
emh203 0:3d9c67d97d6f 292 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 293 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 294
emh203 0:3d9c67d97d6f 295 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 296 acc0 = 0;
emh203 0:3d9c67d97d6f 297 acc1 = 0;
emh203 0:3d9c67d97d6f 298 acc2 = 0;
emh203 0:3d9c67d97d6f 299 acc3 = 0;
emh203 0:3d9c67d97d6f 300
emh203 0:3d9c67d97d6f 301 /* Read two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 302 x1 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 303
emh203 0:3d9c67d97d6f 304 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 305 x2 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 306
emh203 0:3d9c67d97d6f 307 tapCnt = (srcBLen) >> 2u;
emh203 0:3d9c67d97d6f 308
emh203 0:3d9c67d97d6f 309 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 310 {
emh203 0:3d9c67d97d6f 311
emh203 0:3d9c67d97d6f 312 /* Read four samples from smaller buffer */
emh203 0:3d9c67d97d6f 313 y1 = _SIMD32_OFFSET(pScr2);
emh203 0:3d9c67d97d6f 314
emh203 0:3d9c67d97d6f 315 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 316 acc0 = __SMLAD(x1, y1, acc0);
emh203 0:3d9c67d97d6f 317 acc2 = __SMLAD(x2, y1, acc2);
emh203 0:3d9c67d97d6f 318
emh203 0:3d9c67d97d6f 319 /* pack input data */
emh203 0:3d9c67d97d6f 320 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 321 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 322 #else
emh203 0:3d9c67d97d6f 323 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 324 #endif
emh203 0:3d9c67d97d6f 325
emh203 0:3d9c67d97d6f 326 /* multiply and accumlate */
emh203 0:3d9c67d97d6f 327 acc1 = __SMLADX(x3, y1, acc1);
emh203 0:3d9c67d97d6f 328
emh203 0:3d9c67d97d6f 329 /* Read next two samples from scratch1 buffer */
emh203 0:3d9c67d97d6f 330 x1 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 331
emh203 0:3d9c67d97d6f 332 /* pack input data */
emh203 0:3d9c67d97d6f 333 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 334 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 335 #else
emh203 0:3d9c67d97d6f 336 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 337 #endif
emh203 0:3d9c67d97d6f 338
emh203 0:3d9c67d97d6f 339 acc3 = __SMLADX(x3, y1, acc3);
emh203 0:3d9c67d97d6f 340
emh203 0:3d9c67d97d6f 341 /* Read four samples from smaller buffer */
emh203 0:3d9c67d97d6f 342 y1 = _SIMD32_OFFSET(pScr2 + 2u);
emh203 0:3d9c67d97d6f 343
emh203 0:3d9c67d97d6f 344 acc0 = __SMLAD(x2, y1, acc0);
emh203 0:3d9c67d97d6f 345
emh203 0:3d9c67d97d6f 346 acc2 = __SMLAD(x1, y1, acc2);
emh203 0:3d9c67d97d6f 347
emh203 0:3d9c67d97d6f 348 acc1 = __SMLADX(x3, y1, acc1);
emh203 0:3d9c67d97d6f 349
emh203 0:3d9c67d97d6f 350 x2 = *__SIMD32(pScr1)++;
emh203 0:3d9c67d97d6f 351
emh203 0:3d9c67d97d6f 352 #ifndef ARM_MATH_BIG_ENDIAN
emh203 0:3d9c67d97d6f 353 x3 = __PKHBT(x2, x1, 0);
emh203 0:3d9c67d97d6f 354 #else
emh203 0:3d9c67d97d6f 355 x3 = __PKHBT(x1, x2, 0);
emh203 0:3d9c67d97d6f 356 #endif
emh203 0:3d9c67d97d6f 357
emh203 0:3d9c67d97d6f 358 acc3 = __SMLADX(x3, y1, acc3);
emh203 0:3d9c67d97d6f 359
emh203 0:3d9c67d97d6f 360 pScr2 += 4u;
emh203 0:3d9c67d97d6f 361
emh203 0:3d9c67d97d6f 362
emh203 0:3d9c67d97d6f 363 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 364 tapCnt--;
emh203 0:3d9c67d97d6f 365 }
emh203 0:3d9c67d97d6f 366
emh203 0:3d9c67d97d6f 367
emh203 0:3d9c67d97d6f 368
emh203 0:3d9c67d97d6f 369 /* Update scratch pointer for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 370 pScr1 -= 4u;
emh203 0:3d9c67d97d6f 371
emh203 0:3d9c67d97d6f 372
emh203 0:3d9c67d97d6f 373 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 374 tapCnt = (srcBLen) & 3u;
emh203 0:3d9c67d97d6f 375
emh203 0:3d9c67d97d6f 376 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 377 {
emh203 0:3d9c67d97d6f 378
emh203 0:3d9c67d97d6f 379 /* accumlate the results */
emh203 0:3d9c67d97d6f 380 acc0 += (*pScr1++ * *pScr2);
emh203 0:3d9c67d97d6f 381 acc1 += (*pScr1++ * *pScr2);
emh203 0:3d9c67d97d6f 382 acc2 += (*pScr1++ * *pScr2);
emh203 0:3d9c67d97d6f 383 acc3 += (*pScr1++ * *pScr2++);
emh203 0:3d9c67d97d6f 384
emh203 0:3d9c67d97d6f 385 pScr1 -= 3u;
emh203 0:3d9c67d97d6f 386
emh203 0:3d9c67d97d6f 387 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 388 tapCnt--;
emh203 0:3d9c67d97d6f 389 }
emh203 0:3d9c67d97d6f 390
emh203 0:3d9c67d97d6f 391 blkCnt--;
emh203 0:3d9c67d97d6f 392
emh203 0:3d9c67d97d6f 393 /* Store the result in the accumulator in the destination buffer. */
emh203 0:3d9c67d97d6f 394 *pOut = (q7_t) (__SSAT(acc0 >> 7u, 8));
emh203 0:3d9c67d97d6f 395 pOut += inc;
emh203 0:3d9c67d97d6f 396 *pOut = (q7_t) (__SSAT(acc1 >> 7u, 8));
emh203 0:3d9c67d97d6f 397 pOut += inc;
emh203 0:3d9c67d97d6f 398 *pOut = (q7_t) (__SSAT(acc2 >> 7u, 8));
emh203 0:3d9c67d97d6f 399 pOut += inc;
emh203 0:3d9c67d97d6f 400 *pOut = (q7_t) (__SSAT(acc3 >> 7u, 8));
emh203 0:3d9c67d97d6f 401 pOut += inc;
emh203 0:3d9c67d97d6f 402
emh203 0:3d9c67d97d6f 403 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 404 pScr2 = py;
emh203 0:3d9c67d97d6f 405
emh203 0:3d9c67d97d6f 406 pScratch1 += 4u;
emh203 0:3d9c67d97d6f 407
emh203 0:3d9c67d97d6f 408 }
emh203 0:3d9c67d97d6f 409
emh203 0:3d9c67d97d6f 410
emh203 0:3d9c67d97d6f 411 blkCnt = (srcALen + srcBLen - 1u) & 0x3;
emh203 0:3d9c67d97d6f 412
emh203 0:3d9c67d97d6f 413 /* Calculate correlation for remaining samples of Bigger length sequence */
emh203 0:3d9c67d97d6f 414 while(blkCnt > 0)
emh203 0:3d9c67d97d6f 415 {
emh203 0:3d9c67d97d6f 416 /* Initialze temporary scratch pointer as scratch1 */
emh203 0:3d9c67d97d6f 417 pScr1 = pScratch1;
emh203 0:3d9c67d97d6f 418
emh203 0:3d9c67d97d6f 419 /* Clear Accumlators */
emh203 0:3d9c67d97d6f 420 acc0 = 0;
emh203 0:3d9c67d97d6f 421
emh203 0:3d9c67d97d6f 422 tapCnt = (srcBLen) >> 1u;
emh203 0:3d9c67d97d6f 423
emh203 0:3d9c67d97d6f 424 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 425 {
emh203 0:3d9c67d97d6f 426 acc0 += (*pScr1++ * *pScr2++);
emh203 0:3d9c67d97d6f 427 acc0 += (*pScr1++ * *pScr2++);
emh203 0:3d9c67d97d6f 428
emh203 0:3d9c67d97d6f 429 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 430 tapCnt--;
emh203 0:3d9c67d97d6f 431 }
emh203 0:3d9c67d97d6f 432
emh203 0:3d9c67d97d6f 433 tapCnt = (srcBLen) & 1u;
emh203 0:3d9c67d97d6f 434
emh203 0:3d9c67d97d6f 435 /* apply same above for remaining samples of smaller length sequence */
emh203 0:3d9c67d97d6f 436 while(tapCnt > 0u)
emh203 0:3d9c67d97d6f 437 {
emh203 0:3d9c67d97d6f 438
emh203 0:3d9c67d97d6f 439 /* accumlate the results */
emh203 0:3d9c67d97d6f 440 acc0 += (*pScr1++ * *pScr2++);
emh203 0:3d9c67d97d6f 441
emh203 0:3d9c67d97d6f 442 /* Decrement the loop counter */
emh203 0:3d9c67d97d6f 443 tapCnt--;
emh203 0:3d9c67d97d6f 444 }
emh203 0:3d9c67d97d6f 445
emh203 0:3d9c67d97d6f 446 blkCnt--;
emh203 0:3d9c67d97d6f 447
emh203 0:3d9c67d97d6f 448 /* Store the result in the accumulator in the destination buffer. */
emh203 0:3d9c67d97d6f 449 *pOut = (q7_t) (__SSAT(acc0 >> 7u, 8));
emh203 0:3d9c67d97d6f 450
emh203 0:3d9c67d97d6f 451 pOut += inc;
emh203 0:3d9c67d97d6f 452
emh203 0:3d9c67d97d6f 453 /* Initialization of inputB pointer */
emh203 0:3d9c67d97d6f 454 pScr2 = py;
emh203 0:3d9c67d97d6f 455
emh203 0:3d9c67d97d6f 456 pScratch1 += 1u;
emh203 0:3d9c67d97d6f 457
emh203 0:3d9c67d97d6f 458 }
emh203 0:3d9c67d97d6f 459
emh203 0:3d9c67d97d6f 460 }
emh203 0:3d9c67d97d6f 461
emh203 0:3d9c67d97d6f 462 /**
emh203 0:3d9c67d97d6f 463 * @} end of Corr group
emh203 0:3d9c67d97d6f 464 */