mbed official / mbed-dsp

CMSIS DSP library

Dependents:   performance_timer Surfboard_ gps2rtty Capstone ... more

Legacy Warning

This is an mbed 2 library. To learn more about mbed OS 5, visit the docs.

cmsis_dsp/FilteringFunctions/arm_correlate_fast_q31.c@5:3762170b6d4d, 2015-11-20 (annotated)

Committer:
mbed_official
Date:
Fri Nov 20 08:45:18 2015 +0000
Revision:
5:3762170b6d4d
Parent:
3:7a284390b0ce 
Synchronized with git revision 2eb940b9a73af188d3004a2575fdfbb05febe62b



Full URL: https://github.com/mbedmicro/mbed/commit/2eb940b9a73af188d3004a2575fdfbb05febe62b/



Added option to build rpc library. closes #1426

Who changed what in which revision?

UserRevisionLine numberNew contents of line
emilmont 1:fdd22bb7aa52 1 /* ----------------------------------------------------------------------
mbed_official 5:3762170b6d4d 2 * Copyright (C) 2010-2014 ARM Limited. All rights reserved.
emilmont 1:fdd22bb7aa52 3 *
mbed_official 5:3762170b6d4d 4 * $Date: 19. March 2015
mbed_official 5:3762170b6d4d 5 * $Revision: V.1.4.5
emilmont 1:fdd22bb7aa52 6 *
emilmont 2:da51fb522205 7 * Project: CMSIS DSP Library
emilmont 2:da51fb522205 8 * Title: arm_correlate_fast_q31.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: Fast Q31 Correlation.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3
mbed_official 3:7a284390b0ce 13 *
mbed_official 3:7a284390b0ce 14 * Redistribution and use in source and binary forms, with or without
mbed_official 3:7a284390b0ce 15 * modification, are permitted provided that the following conditions
mbed_official 3:7a284390b0ce 16 * are met:
mbed_official 3:7a284390b0ce 17 * - Redistributions of source code must retain the above copyright
mbed_official 3:7a284390b0ce 18 * notice, this list of conditions and the following disclaimer.
mbed_official 3:7a284390b0ce 19 * - Redistributions in binary form must reproduce the above copyright
mbed_official 3:7a284390b0ce 20 * notice, this list of conditions and the following disclaimer in
mbed_official 3:7a284390b0ce 21 * the documentation and/or other materials provided with the
mbed_official 3:7a284390b0ce 22 * distribution.
mbed_official 3:7a284390b0ce 23 * - Neither the name of ARM LIMITED nor the names of its contributors
mbed_official 3:7a284390b0ce 24 * may be used to endorse or promote products derived from this
mbed_official 3:7a284390b0ce 25 * software without specific prior written permission.
mbed_official 3:7a284390b0ce 26 *
mbed_official 3:7a284390b0ce 27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
mbed_official 3:7a284390b0ce 28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
mbed_official 3:7a284390b0ce 29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
mbed_official 3:7a284390b0ce 30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
mbed_official 3:7a284390b0ce 31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
mbed_official 3:7a284390b0ce 32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
mbed_official 3:7a284390b0ce 33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
mbed_official 3:7a284390b0ce 34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
mbed_official 3:7a284390b0ce 35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
mbed_official 3:7a284390b0ce 36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
mbed_official 3:7a284390b0ce 37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
mbed_official 3:7a284390b0ce 38 * POSSIBILITY OF SUCH DAMAGE.
emilmont 1:fdd22bb7aa52 39 * -------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 40
emilmont 1:fdd22bb7aa52 41 #include "arm_math.h"
emilmont 1:fdd22bb7aa52 42
emilmont 1:fdd22bb7aa52 43 /**
emilmont 1:fdd22bb7aa52 44 * @ingroup groupFilters
emilmont 1:fdd22bb7aa52 45 */
emilmont 1:fdd22bb7aa52 46
emilmont 1:fdd22bb7aa52 47 /**
emilmont 1:fdd22bb7aa52 48 * @addtogroup Corr
emilmont 1:fdd22bb7aa52 49 * @{
emilmont 1:fdd22bb7aa52 50 */
emilmont 1:fdd22bb7aa52 51
emilmont 1:fdd22bb7aa52 52 /**
emilmont 1:fdd22bb7aa52 53 * @brief Correlation of Q31 sequences (fast version) for Cortex-M3 and Cortex-M4.
emilmont 1:fdd22bb7aa52 54 * @param[in] *pSrcA points to the first input sequence.
emilmont 1:fdd22bb7aa52 55 * @param[in] srcALen length of the first input sequence.
emilmont 1:fdd22bb7aa52 56 * @param[in] *pSrcB points to the second input sequence.
emilmont 1:fdd22bb7aa52 57 * @param[in] srcBLen length of the second input sequence.
emilmont 1:fdd22bb7aa52 58 * @param[out] *pDst points to the location where the output result is written. Length 2 * max(srcALen, srcBLen) - 1.
emilmont 1:fdd22bb7aa52 59 * @return none.
emilmont 1:fdd22bb7aa52 60 *
emilmont 1:fdd22bb7aa52 61 * @details
emilmont 1:fdd22bb7aa52 62 * <b>Scaling and Overflow Behavior:</b>
emilmont 1:fdd22bb7aa52 63 *
emilmont 1:fdd22bb7aa52 64 * \par
emilmont 1:fdd22bb7aa52 65 * This function is optimized for speed at the expense of fixed-point precision and overflow protection.
emilmont 1:fdd22bb7aa52 66 * The result of each 1.31 x 1.31 multiplication is truncated to 2.30 format.
emilmont 1:fdd22bb7aa52 67 * These intermediate results are accumulated in a 32-bit register in 2.30 format.
emilmont 1:fdd22bb7aa52 68 * Finally, the accumulator is saturated and converted to a 1.31 result.
emilmont 1:fdd22bb7aa52 69 *
emilmont 1:fdd22bb7aa52 70 * \par
emilmont 1:fdd22bb7aa52 71 * The fast version has the same overflow behavior as the standard version but provides less precision since it discards the low 32 bits of each multiplication result.
emilmont 1:fdd22bb7aa52 72 * In order to avoid overflows completely the input signals must be scaled down.
emilmont 1:fdd22bb7aa52 73 * The input signals should be scaled down to avoid intermediate overflows.
emilmont 1:fdd22bb7aa52 74 * Scale down one of the inputs by 1/min(srcALen, srcBLen)to avoid overflows since a
emilmont 1:fdd22bb7aa52 75 * maximum of min(srcALen, srcBLen) number of additions is carried internally.
emilmont 1:fdd22bb7aa52 76 *
emilmont 1:fdd22bb7aa52 77 * \par
emilmont 1:fdd22bb7aa52 78 * See <code>arm_correlate_q31()</code> for a slower implementation of this function which uses 64-bit accumulation to provide higher precision.
emilmont 1:fdd22bb7aa52 79 */
emilmont 1:fdd22bb7aa52 80
emilmont 1:fdd22bb7aa52 81 void arm_correlate_fast_q31(
emilmont 1:fdd22bb7aa52 82 q31_t * pSrcA,
emilmont 1:fdd22bb7aa52 83 uint32_t srcALen,
emilmont 1:fdd22bb7aa52 84 q31_t * pSrcB,
emilmont 1:fdd22bb7aa52 85 uint32_t srcBLen,
emilmont 1:fdd22bb7aa52 86 q31_t * pDst)
emilmont 1:fdd22bb7aa52 87 {
emilmont 1:fdd22bb7aa52 88 q31_t *pIn1; /* inputA pointer */
emilmont 1:fdd22bb7aa52 89 q31_t *pIn2; /* inputB pointer */
emilmont 1:fdd22bb7aa52 90 q31_t *pOut = pDst; /* output pointer */
emilmont 1:fdd22bb7aa52 91 q31_t *px; /* Intermediate inputA pointer */
emilmont 1:fdd22bb7aa52 92 q31_t *py; /* Intermediate inputB pointer */
emilmont 1:fdd22bb7aa52 93 q31_t *pSrc1; /* Intermediate pointers */
emilmont 1:fdd22bb7aa52 94 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */
emilmont 1:fdd22bb7aa52 95 q31_t x0, x1, x2, x3, c0; /* temporary variables for holding input and coefficient values */
emilmont 1:fdd22bb7aa52 96 uint32_t j, k = 0u, count, blkCnt, outBlockSize, blockSize1, blockSize2, blockSize3; /* loop counter */
emilmont 1:fdd22bb7aa52 97 int32_t inc = 1; /* Destination address modifier */
emilmont 1:fdd22bb7aa52 98
emilmont 1:fdd22bb7aa52 99
emilmont 1:fdd22bb7aa52 100 /* The algorithm implementation is based on the lengths of the inputs. */
emilmont 1:fdd22bb7aa52 101 /* srcB is always made to slide across srcA. */
emilmont 1:fdd22bb7aa52 102 /* So srcBLen is always considered as shorter or equal to srcALen */
emilmont 1:fdd22bb7aa52 103 if(srcALen >= srcBLen)
emilmont 1:fdd22bb7aa52 104 {
emilmont 1:fdd22bb7aa52 105 /* Initialization of inputA pointer */
emilmont 1:fdd22bb7aa52 106 pIn1 = (pSrcA);
emilmont 1:fdd22bb7aa52 107
emilmont 1:fdd22bb7aa52 108 /* Initialization of inputB pointer */
emilmont 1:fdd22bb7aa52 109 pIn2 = (pSrcB);
emilmont 1:fdd22bb7aa52 110
emilmont 1:fdd22bb7aa52 111 /* Number of output samples is calculated */
emilmont 1:fdd22bb7aa52 112 outBlockSize = (2u * srcALen) - 1u;
emilmont 1:fdd22bb7aa52 113
emilmont 1:fdd22bb7aa52 114 /* When srcALen > srcBLen, zero padding is done to srcB
emilmont 1:fdd22bb7aa52 115 * to make their lengths equal.
emilmont 1:fdd22bb7aa52 116 * Instead, (outBlockSize - (srcALen + srcBLen - 1))
emilmont 1:fdd22bb7aa52 117 * number of output samples are made zero */
emilmont 1:fdd22bb7aa52 118 j = outBlockSize - (srcALen + (srcBLen - 1u));
emilmont 1:fdd22bb7aa52 119
emilmont 1:fdd22bb7aa52 120 /* Updating the pointer position to non zero value */
emilmont 1:fdd22bb7aa52 121 pOut += j;
emilmont 1:fdd22bb7aa52 122
emilmont 1:fdd22bb7aa52 123 }
emilmont 1:fdd22bb7aa52 124 else
emilmont 1:fdd22bb7aa52 125 {
emilmont 1:fdd22bb7aa52 126 /* Initialization of inputA pointer */
emilmont 1:fdd22bb7aa52 127 pIn1 = (pSrcB);
emilmont 1:fdd22bb7aa52 128
emilmont 1:fdd22bb7aa52 129 /* Initialization of inputB pointer */
emilmont 1:fdd22bb7aa52 130 pIn2 = (pSrcA);
emilmont 1:fdd22bb7aa52 131
emilmont 1:fdd22bb7aa52 132 /* srcBLen is always considered as shorter or equal to srcALen */
emilmont 1:fdd22bb7aa52 133 j = srcBLen;
emilmont 1:fdd22bb7aa52 134 srcBLen = srcALen;
emilmont 1:fdd22bb7aa52 135 srcALen = j;
emilmont 1:fdd22bb7aa52 136
emilmont 1:fdd22bb7aa52 137 /* CORR(x, y) = Reverse order(CORR(y, x)) */
emilmont 1:fdd22bb7aa52 138 /* Hence set the destination pointer to point to the last output sample */
emilmont 1:fdd22bb7aa52 139 pOut = pDst + ((srcALen + srcBLen) - 2u);
emilmont 1:fdd22bb7aa52 140
emilmont 1:fdd22bb7aa52 141 /* Destination address modifier is set to -1 */
emilmont 1:fdd22bb7aa52 142 inc = -1;
emilmont 1:fdd22bb7aa52 143
emilmont 1:fdd22bb7aa52 144 }
emilmont 1:fdd22bb7aa52 145
emilmont 1:fdd22bb7aa52 146 /* The function is internally
emilmont 1:fdd22bb7aa52 147 * divided into three parts according to the number of multiplications that has to be
emilmont 1:fdd22bb7aa52 148 * taken place between inputA samples and inputB samples. In the first part of the
emilmont 1:fdd22bb7aa52 149 * algorithm, the multiplications increase by one for every iteration.
emilmont 1:fdd22bb7aa52 150 * In the second part of the algorithm, srcBLen number of multiplications are done.
emilmont 1:fdd22bb7aa52 151 * In the third part of the algorithm, the multiplications decrease by one
emilmont 1:fdd22bb7aa52 152 * for every iteration.*/
emilmont 1:fdd22bb7aa52 153 /* The algorithm is implemented in three stages.
emilmont 1:fdd22bb7aa52 154 * The loop counters of each stage is initiated here. */
emilmont 1:fdd22bb7aa52 155 blockSize1 = srcBLen - 1u;
emilmont 1:fdd22bb7aa52 156 blockSize2 = srcALen - (srcBLen - 1u);
emilmont 1:fdd22bb7aa52 157 blockSize3 = blockSize1;
emilmont 1:fdd22bb7aa52 158
emilmont 1:fdd22bb7aa52 159 /* --------------------------
emilmont 1:fdd22bb7aa52 160 * Initializations of stage1
emilmont 1:fdd22bb7aa52 161 * -------------------------*/
emilmont 1:fdd22bb7aa52 162
emilmont 1:fdd22bb7aa52 163 /* sum = x[0] * y[srcBlen - 1]
emilmont 1:fdd22bb7aa52 164 * sum = x[0] * y[srcBlen - 2] + x[1] * y[srcBlen - 1]
emilmont 1:fdd22bb7aa52 165 * ....
emilmont 1:fdd22bb7aa52 166 * sum = x[0] * y[0] + x[1] * y[1] +...+ x[srcBLen - 1] * y[srcBLen - 1]
emilmont 1:fdd22bb7aa52 167 */
emilmont 1:fdd22bb7aa52 168
emilmont 1:fdd22bb7aa52 169 /* In this stage the MAC operations are increased by 1 for every iteration.
emilmont 1:fdd22bb7aa52 170 The count variable holds the number of MAC operations performed */
emilmont 1:fdd22bb7aa52 171 count = 1u;
emilmont 1:fdd22bb7aa52 172
emilmont 1:fdd22bb7aa52 173 /* Working pointer of inputA */
emilmont 1:fdd22bb7aa52 174 px = pIn1;
emilmont 1:fdd22bb7aa52 175
emilmont 1:fdd22bb7aa52 176 /* Working pointer of inputB */
emilmont 1:fdd22bb7aa52 177 pSrc1 = pIn2 + (srcBLen - 1u);
emilmont 1:fdd22bb7aa52 178 py = pSrc1;
emilmont 1:fdd22bb7aa52 179
emilmont 1:fdd22bb7aa52 180 /* ------------------------
emilmont 1:fdd22bb7aa52 181 * Stage1 process
emilmont 1:fdd22bb7aa52 182 * ----------------------*/
emilmont 1:fdd22bb7aa52 183
emilmont 1:fdd22bb7aa52 184 /* The first stage starts here */
emilmont 1:fdd22bb7aa52 185 while(blockSize1 > 0u)
emilmont 1:fdd22bb7aa52 186 {
emilmont 1:fdd22bb7aa52 187 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 188 sum = 0;
emilmont 1:fdd22bb7aa52 189
emilmont 1:fdd22bb7aa52 190 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 191 k = count >> 2;
emilmont 1:fdd22bb7aa52 192
emilmont 1:fdd22bb7aa52 193 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 194 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 195 while(k > 0u)
emilmont 1:fdd22bb7aa52 196 {
emilmont 1:fdd22bb7aa52 197 /* x[0] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 198 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 199 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 200 /* x[1] * y[srcBLen - 3] */
emilmont 1:fdd22bb7aa52 201 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 202 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 203 /* x[2] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 204 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 205 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 206 /* x[3] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 207 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 208 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 209
emilmont 1:fdd22bb7aa52 210 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 211 k--;
emilmont 1:fdd22bb7aa52 212 }
emilmont 1:fdd22bb7aa52 213
emilmont 1:fdd22bb7aa52 214 /* If the count is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 215 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 216 k = count % 0x4u;
emilmont 1:fdd22bb7aa52 217
emilmont 1:fdd22bb7aa52 218 while(k > 0u)
emilmont 1:fdd22bb7aa52 219 {
emilmont 1:fdd22bb7aa52 220 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 221 /* x[0] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 222 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 223 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 224
emilmont 1:fdd22bb7aa52 225 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 226 k--;
emilmont 1:fdd22bb7aa52 227 }
emilmont 1:fdd22bb7aa52 228
emilmont 1:fdd22bb7aa52 229 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 230 *pOut = sum << 1;
emilmont 1:fdd22bb7aa52 231 /* Destination pointer is updated according to the address modifier, inc */
emilmont 1:fdd22bb7aa52 232 pOut += inc;
emilmont 1:fdd22bb7aa52 233
emilmont 1:fdd22bb7aa52 234 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 235 py = pSrc1 - count;
emilmont 1:fdd22bb7aa52 236 px = pIn1;
emilmont 1:fdd22bb7aa52 237
emilmont 1:fdd22bb7aa52 238 /* Increment the MAC count */
emilmont 1:fdd22bb7aa52 239 count++;
emilmont 1:fdd22bb7aa52 240
emilmont 1:fdd22bb7aa52 241 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 242 blockSize1--;
emilmont 1:fdd22bb7aa52 243 }
emilmont 1:fdd22bb7aa52 244
emilmont 1:fdd22bb7aa52 245 /* --------------------------
emilmont 1:fdd22bb7aa52 246 * Initializations of stage2
emilmont 1:fdd22bb7aa52 247 * ------------------------*/
emilmont 1:fdd22bb7aa52 248
emilmont 1:fdd22bb7aa52 249 /* sum = x[0] * y[0] + x[1] * y[1] +...+ x[srcBLen-1] * y[srcBLen-1]
emilmont 1:fdd22bb7aa52 250 * sum = x[1] * y[0] + x[2] * y[1] +...+ x[srcBLen] * y[srcBLen-1]
emilmont 1:fdd22bb7aa52 251 * ....
emilmont 1:fdd22bb7aa52 252 * sum = x[srcALen-srcBLen-2] * y[0] + x[srcALen-srcBLen-1] * y[1] +...+ x[srcALen-1] * y[srcBLen-1]
emilmont 1:fdd22bb7aa52 253 */
emilmont 1:fdd22bb7aa52 254
emilmont 1:fdd22bb7aa52 255 /* Working pointer of inputA */
emilmont 1:fdd22bb7aa52 256 px = pIn1;
emilmont 1:fdd22bb7aa52 257
emilmont 1:fdd22bb7aa52 258 /* Working pointer of inputB */
emilmont 1:fdd22bb7aa52 259 py = pIn2;
emilmont 1:fdd22bb7aa52 260
emilmont 1:fdd22bb7aa52 261 /* count is index by which the pointer pIn1 to be incremented */
emilmont 1:fdd22bb7aa52 262 count = 0u;
emilmont 1:fdd22bb7aa52 263
emilmont 1:fdd22bb7aa52 264 /* -------------------
emilmont 1:fdd22bb7aa52 265 * Stage2 process
emilmont 1:fdd22bb7aa52 266 * ------------------*/
emilmont 1:fdd22bb7aa52 267
emilmont 1:fdd22bb7aa52 268 /* Stage2 depends on srcBLen as in this stage srcBLen number of MACS are performed.
emilmont 1:fdd22bb7aa52 269 * So, to loop unroll over blockSize2,
emilmont 1:fdd22bb7aa52 270 * srcBLen should be greater than or equal to 4 */
emilmont 1:fdd22bb7aa52 271 if(srcBLen >= 4u)
emilmont 1:fdd22bb7aa52 272 {
emilmont 1:fdd22bb7aa52 273 /* Loop unroll over blockSize2, by 4 */
emilmont 1:fdd22bb7aa52 274 blkCnt = blockSize2 >> 2u;
emilmont 1:fdd22bb7aa52 275
emilmont 1:fdd22bb7aa52 276 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 277 {
emilmont 1:fdd22bb7aa52 278 /* Set all accumulators to zero */
emilmont 1:fdd22bb7aa52 279 acc0 = 0;
emilmont 1:fdd22bb7aa52 280 acc1 = 0;
emilmont 1:fdd22bb7aa52 281 acc2 = 0;
emilmont 1:fdd22bb7aa52 282 acc3 = 0;
emilmont 1:fdd22bb7aa52 283
emilmont 1:fdd22bb7aa52 284 /* read x[0], x[1], x[2] samples */
emilmont 1:fdd22bb7aa52 285 x0 = *(px++);
emilmont 1:fdd22bb7aa52 286 x1 = *(px++);
emilmont 1:fdd22bb7aa52 287 x2 = *(px++);
emilmont 1:fdd22bb7aa52 288
emilmont 1:fdd22bb7aa52 289 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 290 k = srcBLen >> 2u;
emilmont 1:fdd22bb7aa52 291
emilmont 1:fdd22bb7aa52 292 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 293 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 294 do
emilmont 1:fdd22bb7aa52 295 {
emilmont 1:fdd22bb7aa52 296 /* Read y[0] sample */
emilmont 1:fdd22bb7aa52 297 c0 = *(py++);
emilmont 1:fdd22bb7aa52 298
emilmont 1:fdd22bb7aa52 299 /* Read x[3] sample */
emilmont 1:fdd22bb7aa52 300 x3 = *(px++);
emilmont 1:fdd22bb7aa52 301
emilmont 1:fdd22bb7aa52 302 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 303 /* acc0 += x[0] * y[0] */
emilmont 1:fdd22bb7aa52 304 acc0 = (q31_t) ((((q63_t) acc0 << 32) + ((q63_t) x0 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 305 /* acc1 += x[1] * y[0] */
emilmont 1:fdd22bb7aa52 306 acc1 = (q31_t) ((((q63_t) acc1 << 32) + ((q63_t) x1 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 307 /* acc2 += x[2] * y[0] */
emilmont 1:fdd22bb7aa52 308 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 309 /* acc3 += x[3] * y[0] */
emilmont 1:fdd22bb7aa52 310 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 311
emilmont 1:fdd22bb7aa52 312 /* Read y[1] sample */
emilmont 1:fdd22bb7aa52 313 c0 = *(py++);
emilmont 1:fdd22bb7aa52 314
emilmont 1:fdd22bb7aa52 315 /* Read x[4] sample */
emilmont 1:fdd22bb7aa52 316 x0 = *(px++);
emilmont 1:fdd22bb7aa52 317
emilmont 1:fdd22bb7aa52 318 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 319 /* acc0 += x[1] * y[1] */
emilmont 1:fdd22bb7aa52 320 acc0 = (q31_t) ((((q63_t) acc0 << 32) + ((q63_t) x1 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 321 /* acc1 += x[2] * y[1] */
emilmont 1:fdd22bb7aa52 322 acc1 = (q31_t) ((((q63_t) acc1 << 32) + ((q63_t) x2 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 323 /* acc2 += x[3] * y[1] */
emilmont 1:fdd22bb7aa52 324 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x3 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 325 /* acc3 += x[4] * y[1] */
emilmont 1:fdd22bb7aa52 326 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 327
emilmont 1:fdd22bb7aa52 328 /* Read y[2] sample */
emilmont 1:fdd22bb7aa52 329 c0 = *(py++);
emilmont 1:fdd22bb7aa52 330
emilmont 1:fdd22bb7aa52 331 /* Read x[5] sample */
emilmont 1:fdd22bb7aa52 332 x1 = *(px++);
emilmont 1:fdd22bb7aa52 333
emilmont 1:fdd22bb7aa52 334 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 335 /* acc0 += x[2] * y[2] */
emilmont 1:fdd22bb7aa52 336 acc0 = (q31_t) ((((q63_t) acc0 << 32) + ((q63_t) x2 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 337 /* acc1 += x[3] * y[2] */
emilmont 1:fdd22bb7aa52 338 acc1 = (q31_t) ((((q63_t) acc1 << 32) + ((q63_t) x3 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 339 /* acc2 += x[4] * y[2] */
emilmont 1:fdd22bb7aa52 340 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x0 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 341 /* acc3 += x[5] * y[2] */
emilmont 1:fdd22bb7aa52 342 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 343
emilmont 1:fdd22bb7aa52 344 /* Read y[3] sample */
emilmont 1:fdd22bb7aa52 345 c0 = *(py++);
emilmont 1:fdd22bb7aa52 346
emilmont 1:fdd22bb7aa52 347 /* Read x[6] sample */
emilmont 1:fdd22bb7aa52 348 x2 = *(px++);
emilmont 1:fdd22bb7aa52 349
emilmont 1:fdd22bb7aa52 350 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 351 /* acc0 += x[3] * y[3] */
emilmont 1:fdd22bb7aa52 352 acc0 = (q31_t) ((((q63_t) acc0 << 32) + ((q63_t) x3 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 353 /* acc1 += x[4] * y[3] */
emilmont 1:fdd22bb7aa52 354 acc1 = (q31_t) ((((q63_t) acc1 << 32) + ((q63_t) x0 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 355 /* acc2 += x[5] * y[3] */
emilmont 1:fdd22bb7aa52 356 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x1 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 357 /* acc3 += x[6] * y[3] */
emilmont 1:fdd22bb7aa52 358 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x2 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 359
emilmont 1:fdd22bb7aa52 360
emilmont 1:fdd22bb7aa52 361 } while(--k);
emilmont 1:fdd22bb7aa52 362
emilmont 1:fdd22bb7aa52 363 /* If the srcBLen is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 364 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 365 k = srcBLen % 0x4u;
emilmont 1:fdd22bb7aa52 366
emilmont 1:fdd22bb7aa52 367 while(k > 0u)
emilmont 1:fdd22bb7aa52 368 {
emilmont 1:fdd22bb7aa52 369 /* Read y[4] sample */
emilmont 1:fdd22bb7aa52 370 c0 = *(py++);
emilmont 1:fdd22bb7aa52 371
emilmont 1:fdd22bb7aa52 372 /* Read x[7] sample */
emilmont 1:fdd22bb7aa52 373 x3 = *(px++);
emilmont 1:fdd22bb7aa52 374
emilmont 1:fdd22bb7aa52 375 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 376 /* acc0 += x[4] * y[4] */
emilmont 1:fdd22bb7aa52 377 acc0 = (q31_t) ((((q63_t) acc0 << 32) + ((q63_t) x0 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 378 /* acc1 += x[5] * y[4] */
emilmont 1:fdd22bb7aa52 379 acc1 = (q31_t) ((((q63_t) acc1 << 32) + ((q63_t) x1 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 380 /* acc2 += x[6] * y[4] */
emilmont 1:fdd22bb7aa52 381 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 382 /* acc3 += x[7] * y[4] */
emilmont 1:fdd22bb7aa52 383 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32);
emilmont 1:fdd22bb7aa52 384
emilmont 1:fdd22bb7aa52 385 /* Reuse the present samples for the next MAC */
emilmont 1:fdd22bb7aa52 386 x0 = x1;
emilmont 1:fdd22bb7aa52 387 x1 = x2;
emilmont 1:fdd22bb7aa52 388 x2 = x3;
emilmont 1:fdd22bb7aa52 389
emilmont 1:fdd22bb7aa52 390 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 391 k--;
emilmont 1:fdd22bb7aa52 392 }
emilmont 1:fdd22bb7aa52 393
emilmont 1:fdd22bb7aa52 394 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 395 *pOut = (q31_t) (acc0 << 1);
emilmont 1:fdd22bb7aa52 396 /* Destination pointer is updated according to the address modifier, inc */
emilmont 1:fdd22bb7aa52 397 pOut += inc;
emilmont 1:fdd22bb7aa52 398
emilmont 1:fdd22bb7aa52 399 *pOut = (q31_t) (acc1 << 1);
emilmont 1:fdd22bb7aa52 400 pOut += inc;
emilmont 1:fdd22bb7aa52 401
emilmont 1:fdd22bb7aa52 402 *pOut = (q31_t) (acc2 << 1);
emilmont 1:fdd22bb7aa52 403 pOut += inc;
emilmont 1:fdd22bb7aa52 404
emilmont 1:fdd22bb7aa52 405 *pOut = (q31_t) (acc3 << 1);
emilmont 1:fdd22bb7aa52 406 pOut += inc;
emilmont 1:fdd22bb7aa52 407
emilmont 1:fdd22bb7aa52 408 /* Increment the pointer pIn1 index, count by 4 */
emilmont 1:fdd22bb7aa52 409 count += 4u;
emilmont 1:fdd22bb7aa52 410
emilmont 1:fdd22bb7aa52 411 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 412 px = pIn1 + count;
emilmont 1:fdd22bb7aa52 413 py = pIn2;
emilmont 1:fdd22bb7aa52 414
emilmont 1:fdd22bb7aa52 415
emilmont 1:fdd22bb7aa52 416 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 417 blkCnt--;
emilmont 1:fdd22bb7aa52 418 }
emilmont 1:fdd22bb7aa52 419
emilmont 1:fdd22bb7aa52 420 /* If the blockSize2 is not a multiple of 4, compute any remaining output samples here.
emilmont 1:fdd22bb7aa52 421 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 422 blkCnt = blockSize2 % 0x4u;
emilmont 1:fdd22bb7aa52 423
emilmont 1:fdd22bb7aa52 424 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 425 {
emilmont 1:fdd22bb7aa52 426 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 427 sum = 0;
emilmont 1:fdd22bb7aa52 428
emilmont 1:fdd22bb7aa52 429 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 430 k = srcBLen >> 2u;
emilmont 1:fdd22bb7aa52 431
emilmont 1:fdd22bb7aa52 432 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 433 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 434 while(k > 0u)
emilmont 1:fdd22bb7aa52 435 {
emilmont 1:fdd22bb7aa52 436 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 437 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 438 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 439 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 440 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 441 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 442 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 443 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 444 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 445
emilmont 1:fdd22bb7aa52 446 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 447 k--;
emilmont 1:fdd22bb7aa52 448 }
emilmont 1:fdd22bb7aa52 449
emilmont 1:fdd22bb7aa52 450 /* If the srcBLen is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 451 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 452 k = srcBLen % 0x4u;
emilmont 1:fdd22bb7aa52 453
emilmont 1:fdd22bb7aa52 454 while(k > 0u)
emilmont 1:fdd22bb7aa52 455 {
emilmont 1:fdd22bb7aa52 456 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 457 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 458 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 459
emilmont 1:fdd22bb7aa52 460 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 461 k--;
emilmont 1:fdd22bb7aa52 462 }
emilmont 1:fdd22bb7aa52 463
emilmont 1:fdd22bb7aa52 464 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 465 *pOut = sum << 1;
emilmont 1:fdd22bb7aa52 466 /* Destination pointer is updated according to the address modifier, inc */
emilmont 1:fdd22bb7aa52 467 pOut += inc;
emilmont 1:fdd22bb7aa52 468
emilmont 1:fdd22bb7aa52 469 /* Increment the MAC count */
emilmont 1:fdd22bb7aa52 470 count++;
emilmont 1:fdd22bb7aa52 471
emilmont 1:fdd22bb7aa52 472 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 473 px = pIn1 + count;
emilmont 1:fdd22bb7aa52 474 py = pIn2;
emilmont 1:fdd22bb7aa52 475
emilmont 1:fdd22bb7aa52 476
emilmont 1:fdd22bb7aa52 477 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 478 blkCnt--;
emilmont 1:fdd22bb7aa52 479 }
emilmont 1:fdd22bb7aa52 480 }
emilmont 1:fdd22bb7aa52 481 else
emilmont 1:fdd22bb7aa52 482 {
emilmont 1:fdd22bb7aa52 483 /* If the srcBLen is not a multiple of 4,
emilmont 1:fdd22bb7aa52 484 * the blockSize2 loop cannot be unrolled by 4 */
emilmont 1:fdd22bb7aa52 485 blkCnt = blockSize2;
emilmont 1:fdd22bb7aa52 486
emilmont 1:fdd22bb7aa52 487 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 488 {
emilmont 1:fdd22bb7aa52 489 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 490 sum = 0;
emilmont 1:fdd22bb7aa52 491
emilmont 1:fdd22bb7aa52 492 /* Loop over srcBLen */
emilmont 1:fdd22bb7aa52 493 k = srcBLen;
emilmont 1:fdd22bb7aa52 494
emilmont 1:fdd22bb7aa52 495 while(k > 0u)
emilmont 1:fdd22bb7aa52 496 {
emilmont 1:fdd22bb7aa52 497 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 498 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 499 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 500
emilmont 1:fdd22bb7aa52 501 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 502 k--;
emilmont 1:fdd22bb7aa52 503 }
emilmont 1:fdd22bb7aa52 504
emilmont 1:fdd22bb7aa52 505 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 506 *pOut = sum << 1;
emilmont 1:fdd22bb7aa52 507 /* Destination pointer is updated according to the address modifier, inc */
emilmont 1:fdd22bb7aa52 508 pOut += inc;
emilmont 1:fdd22bb7aa52 509
emilmont 1:fdd22bb7aa52 510 /* Increment the MAC count */
emilmont 1:fdd22bb7aa52 511 count++;
emilmont 1:fdd22bb7aa52 512
emilmont 1:fdd22bb7aa52 513 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 514 px = pIn1 + count;
emilmont 1:fdd22bb7aa52 515 py = pIn2;
emilmont 1:fdd22bb7aa52 516
emilmont 1:fdd22bb7aa52 517 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 518 blkCnt--;
emilmont 1:fdd22bb7aa52 519 }
emilmont 1:fdd22bb7aa52 520 }
emilmont 1:fdd22bb7aa52 521
emilmont 1:fdd22bb7aa52 522 /* --------------------------
emilmont 1:fdd22bb7aa52 523 * Initializations of stage3
emilmont 1:fdd22bb7aa52 524 * -------------------------*/
emilmont 1:fdd22bb7aa52 525
emilmont 1:fdd22bb7aa52 526 /* sum += x[srcALen-srcBLen+1] * y[0] + x[srcALen-srcBLen+2] * y[1] +...+ x[srcALen-1] * y[srcBLen-1]
emilmont 1:fdd22bb7aa52 527 * sum += x[srcALen-srcBLen+2] * y[0] + x[srcALen-srcBLen+3] * y[1] +...+ x[srcALen-1] * y[srcBLen-1]
emilmont 1:fdd22bb7aa52 528 * ....
emilmont 1:fdd22bb7aa52 529 * sum += x[srcALen-2] * y[0] + x[srcALen-1] * y[1]
emilmont 1:fdd22bb7aa52 530 * sum += x[srcALen-1] * y[0]
emilmont 1:fdd22bb7aa52 531 */
emilmont 1:fdd22bb7aa52 532
emilmont 1:fdd22bb7aa52 533 /* In this stage the MAC operations are decreased by 1 for every iteration.
emilmont 1:fdd22bb7aa52 534 The count variable holds the number of MAC operations performed */
emilmont 1:fdd22bb7aa52 535 count = srcBLen - 1u;
emilmont 1:fdd22bb7aa52 536
emilmont 1:fdd22bb7aa52 537 /* Working pointer of inputA */
emilmont 1:fdd22bb7aa52 538 pSrc1 = ((pIn1 + srcALen) - srcBLen) + 1u;
emilmont 1:fdd22bb7aa52 539 px = pSrc1;
emilmont 1:fdd22bb7aa52 540
emilmont 1:fdd22bb7aa52 541 /* Working pointer of inputB */
emilmont 1:fdd22bb7aa52 542 py = pIn2;
emilmont 1:fdd22bb7aa52 543
emilmont 1:fdd22bb7aa52 544 /* -------------------
emilmont 1:fdd22bb7aa52 545 * Stage3 process
emilmont 1:fdd22bb7aa52 546 * ------------------*/
emilmont 1:fdd22bb7aa52 547
emilmont 1:fdd22bb7aa52 548 while(blockSize3 > 0u)
emilmont 1:fdd22bb7aa52 549 {
emilmont 1:fdd22bb7aa52 550 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 551 sum = 0;
emilmont 1:fdd22bb7aa52 552
emilmont 1:fdd22bb7aa52 553 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 554 k = count >> 2u;
emilmont 1:fdd22bb7aa52 555
emilmont 1:fdd22bb7aa52 556 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 557 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 558 while(k > 0u)
emilmont 1:fdd22bb7aa52 559 {
emilmont 1:fdd22bb7aa52 560 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 561 /* sum += x[srcALen - srcBLen + 4] * y[3] */
emilmont 1:fdd22bb7aa52 562 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 563 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 564 /* sum += x[srcALen - srcBLen + 3] * y[2] */
emilmont 1:fdd22bb7aa52 565 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 566 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 567 /* sum += x[srcALen - srcBLen + 2] * y[1] */
emilmont 1:fdd22bb7aa52 568 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 569 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 570 /* sum += x[srcALen - srcBLen + 1] * y[0] */
emilmont 1:fdd22bb7aa52 571 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 572 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 573
emilmont 1:fdd22bb7aa52 574 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 575 k--;
emilmont 1:fdd22bb7aa52 576 }
emilmont 1:fdd22bb7aa52 577
emilmont 1:fdd22bb7aa52 578 /* If the count is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 579 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 580 k = count % 0x4u;
emilmont 1:fdd22bb7aa52 581
emilmont 1:fdd22bb7aa52 582 while(k > 0u)
emilmont 1:fdd22bb7aa52 583 {
emilmont 1:fdd22bb7aa52 584 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 585 sum = (q31_t) ((((q63_t) sum << 32) +
emilmont 1:fdd22bb7aa52 586 ((q63_t) * px++ * (*py++))) >> 32);
emilmont 1:fdd22bb7aa52 587
emilmont 1:fdd22bb7aa52 588 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 589 k--;
emilmont 1:fdd22bb7aa52 590 }
emilmont 1:fdd22bb7aa52 591
emilmont 1:fdd22bb7aa52 592 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 593 *pOut = sum << 1;
emilmont 1:fdd22bb7aa52 594 /* Destination pointer is updated according to the address modifier, inc */
emilmont 1:fdd22bb7aa52 595 pOut += inc;
emilmont 1:fdd22bb7aa52 596
emilmont 1:fdd22bb7aa52 597 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 598 px = ++pSrc1;
emilmont 1:fdd22bb7aa52 599 py = pIn2;
emilmont 1:fdd22bb7aa52 600
emilmont 1:fdd22bb7aa52 601 /* Decrement the MAC count */
emilmont 1:fdd22bb7aa52 602 count--;
emilmont 1:fdd22bb7aa52 603
emilmont 1:fdd22bb7aa52 604 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 605 blockSize3--;
emilmont 1:fdd22bb7aa52 606 }
emilmont 1:fdd22bb7aa52 607
emilmont 1:fdd22bb7aa52 608 }
emilmont 1:fdd22bb7aa52 609
emilmont 1:fdd22bb7aa52 610 /**
emilmont 1:fdd22bb7aa52 611 * @} end of Corr group
emilmont 1:fdd22bb7aa52 612 */