CMSIS DSP library
Dependents: KL25Z_FFT_Demo Hat_Board_v5_1 KL25Z_FFT_Demo_tony KL25Z_FFT_Demo_tony ... more
Fork of mbed-dsp by
cmsis_dsp/FilteringFunctions/arm_fir_fast_q15.c@3:7a284390b0ce, 2013-11-08 (annotated)
- Committer:
- mbed_official
- Date:
- Fri Nov 08 13:45:10 2013 +0000
- Revision:
- 3:7a284390b0ce
- Parent:
- 2:da51fb522205
Synchronized with git revision e69956aba2f68a2a26ac26b051f8d349deaa1ce8
Who changed what in which revision?
User | Revision | Line number | New contents of line |
---|---|---|---|
emilmont | 1:fdd22bb7aa52 | 1 | /* ---------------------------------------------------------------------- |
mbed_official | 3:7a284390b0ce | 2 | * Copyright (C) 2010-2013 ARM Limited. All rights reserved. |
emilmont | 1:fdd22bb7aa52 | 3 | * |
mbed_official | 3:7a284390b0ce | 4 | * $Date: 17. January 2013 |
mbed_official | 3:7a284390b0ce | 5 | * $Revision: V1.4.1 |
emilmont | 1:fdd22bb7aa52 | 6 | * |
emilmont | 2:da51fb522205 | 7 | * Project: CMSIS DSP Library |
emilmont | 1:fdd22bb7aa52 | 8 | * Title: arm_fir_fast_q15.c |
emilmont | 1:fdd22bb7aa52 | 9 | * |
emilmont | 1:fdd22bb7aa52 | 10 | * Description: Q15 Fast FIR filter processing function. |
emilmont | 1:fdd22bb7aa52 | 11 | * |
emilmont | 1:fdd22bb7aa52 | 12 | * Target Processor: Cortex-M4/Cortex-M3 |
emilmont | 1:fdd22bb7aa52 | 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 FIR |
emilmont | 1:fdd22bb7aa52 | 49 | * @{ |
emilmont | 1:fdd22bb7aa52 | 50 | */ |
emilmont | 1:fdd22bb7aa52 | 51 | |
emilmont | 1:fdd22bb7aa52 | 52 | /** |
emilmont | 1:fdd22bb7aa52 | 53 | * @param[in] *S points to an instance of the Q15 FIR filter structure. |
emilmont | 1:fdd22bb7aa52 | 54 | * @param[in] *pSrc points to the block of input data. |
emilmont | 1:fdd22bb7aa52 | 55 | * @param[out] *pDst points to the block of output data. |
emilmont | 1:fdd22bb7aa52 | 56 | * @param[in] blockSize number of samples to process per call. |
emilmont | 1:fdd22bb7aa52 | 57 | * @return none. |
emilmont | 1:fdd22bb7aa52 | 58 | * |
emilmont | 1:fdd22bb7aa52 | 59 | * <b>Scaling and Overflow Behavior:</b> |
emilmont | 1:fdd22bb7aa52 | 60 | * \par |
emilmont | 1:fdd22bb7aa52 | 61 | * This fast version uses a 32-bit accumulator with 2.30 format. |
emilmont | 1:fdd22bb7aa52 | 62 | * The accumulator maintains full precision of the intermediate multiplication results but provides only a single guard bit. |
emilmont | 1:fdd22bb7aa52 | 63 | * Thus, if the accumulator result overflows it wraps around and distorts the result. |
emilmont | 1:fdd22bb7aa52 | 64 | * In order to avoid overflows completely the input signal must be scaled down by log2(numTaps) bits. |
emilmont | 1:fdd22bb7aa52 | 65 | * The 2.30 accumulator is then truncated to 2.15 format and saturated to yield the 1.15 result. |
emilmont | 1:fdd22bb7aa52 | 66 | * |
emilmont | 1:fdd22bb7aa52 | 67 | * \par |
emilmont | 1:fdd22bb7aa52 | 68 | * Refer to the function <code>arm_fir_q15()</code> for a slower implementation of this function which uses 64-bit accumulation to avoid wrap around distortion. Both the slow and the fast versions use the same instance structure. |
emilmont | 1:fdd22bb7aa52 | 69 | * Use the function <code>arm_fir_init_q15()</code> to initialize the filter structure. |
emilmont | 1:fdd22bb7aa52 | 70 | */ |
emilmont | 1:fdd22bb7aa52 | 71 | |
emilmont | 1:fdd22bb7aa52 | 72 | void arm_fir_fast_q15( |
emilmont | 1:fdd22bb7aa52 | 73 | const arm_fir_instance_q15 * S, |
emilmont | 1:fdd22bb7aa52 | 74 | q15_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 75 | q15_t * pDst, |
emilmont | 1:fdd22bb7aa52 | 76 | uint32_t blockSize) |
emilmont | 1:fdd22bb7aa52 | 77 | { |
emilmont | 1:fdd22bb7aa52 | 78 | q15_t *pState = S->pState; /* State pointer */ |
emilmont | 1:fdd22bb7aa52 | 79 | q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */ |
emilmont | 1:fdd22bb7aa52 | 80 | q15_t *pStateCurnt; /* Points to the current sample of the state */ |
emilmont | 1:fdd22bb7aa52 | 81 | q31_t acc0, acc1, acc2, acc3; /* Accumulators */ |
emilmont | 1:fdd22bb7aa52 | 82 | q15_t *pb; /* Temporary pointer for coefficient buffer */ |
emilmont | 1:fdd22bb7aa52 | 83 | q15_t *px; /* Temporary q31 pointer for SIMD state buffer accesses */ |
emilmont | 1:fdd22bb7aa52 | 84 | q31_t x0, x1, x2, c0; /* Temporary variables to hold SIMD state and coefficient values */ |
emilmont | 1:fdd22bb7aa52 | 85 | uint32_t numTaps = S->numTaps; /* Number of taps in the filter */ |
emilmont | 1:fdd22bb7aa52 | 86 | uint32_t tapCnt, blkCnt; /* Loop counters */ |
emilmont | 1:fdd22bb7aa52 | 87 | |
emilmont | 1:fdd22bb7aa52 | 88 | |
emilmont | 1:fdd22bb7aa52 | 89 | /* S->pState points to state array which contains previous frame (numTaps - 1) samples */ |
emilmont | 1:fdd22bb7aa52 | 90 | /* pStateCurnt points to the location where the new input data should be written */ |
emilmont | 1:fdd22bb7aa52 | 91 | pStateCurnt = &(S->pState[(numTaps - 1u)]); |
emilmont | 1:fdd22bb7aa52 | 92 | |
emilmont | 1:fdd22bb7aa52 | 93 | /* Apply loop unrolling and compute 4 output values simultaneously. |
emilmont | 1:fdd22bb7aa52 | 94 | * The variables acc0 ... acc3 hold output values that are being computed: |
emilmont | 1:fdd22bb7aa52 | 95 | * |
emilmont | 1:fdd22bb7aa52 | 96 | * acc0 = b[numTaps-1] * x[n-numTaps-1] + b[numTaps-2] * x[n-numTaps-2] + b[numTaps-3] * x[n-numTaps-3] +...+ b[0] * x[0] |
emilmont | 1:fdd22bb7aa52 | 97 | * acc1 = b[numTaps-1] * x[n-numTaps] + b[numTaps-2] * x[n-numTaps-1] + b[numTaps-3] * x[n-numTaps-2] +...+ b[0] * x[1] |
emilmont | 1:fdd22bb7aa52 | 98 | * acc2 = b[numTaps-1] * x[n-numTaps+1] + b[numTaps-2] * x[n-numTaps] + b[numTaps-3] * x[n-numTaps-1] +...+ b[0] * x[2] |
emilmont | 1:fdd22bb7aa52 | 99 | * acc3 = b[numTaps-1] * x[n-numTaps+2] + b[numTaps-2] * x[n-numTaps+1] + b[numTaps-3] * x[n-numTaps] +...+ b[0] * x[3] |
emilmont | 1:fdd22bb7aa52 | 100 | */ |
emilmont | 1:fdd22bb7aa52 | 101 | |
emilmont | 1:fdd22bb7aa52 | 102 | blkCnt = blockSize >> 2; |
emilmont | 1:fdd22bb7aa52 | 103 | |
emilmont | 1:fdd22bb7aa52 | 104 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
emilmont | 1:fdd22bb7aa52 | 105 | ** a second loop below computes the remaining 1 to 3 samples. */ |
emilmont | 1:fdd22bb7aa52 | 106 | while(blkCnt > 0u) |
emilmont | 1:fdd22bb7aa52 | 107 | { |
emilmont | 1:fdd22bb7aa52 | 108 | /* Copy four new input samples into the state buffer. |
emilmont | 1:fdd22bb7aa52 | 109 | ** Use 32-bit SIMD to move the 16-bit data. Only requires two copies. */ |
emilmont | 1:fdd22bb7aa52 | 110 | *pStateCurnt++ = *pSrc++; |
emilmont | 1:fdd22bb7aa52 | 111 | *pStateCurnt++ = *pSrc++; |
emilmont | 1:fdd22bb7aa52 | 112 | *pStateCurnt++ = *pSrc++; |
emilmont | 1:fdd22bb7aa52 | 113 | *pStateCurnt++ = *pSrc++; |
emilmont | 1:fdd22bb7aa52 | 114 | |
emilmont | 1:fdd22bb7aa52 | 115 | |
emilmont | 1:fdd22bb7aa52 | 116 | /* Set all accumulators to zero */ |
emilmont | 1:fdd22bb7aa52 | 117 | acc0 = 0; |
emilmont | 1:fdd22bb7aa52 | 118 | acc1 = 0; |
emilmont | 1:fdd22bb7aa52 | 119 | acc2 = 0; |
emilmont | 1:fdd22bb7aa52 | 120 | acc3 = 0; |
emilmont | 1:fdd22bb7aa52 | 121 | |
emilmont | 1:fdd22bb7aa52 | 122 | /* Typecast q15_t pointer to q31_t pointer for state reading in q31_t */ |
emilmont | 1:fdd22bb7aa52 | 123 | px = pState; |
emilmont | 1:fdd22bb7aa52 | 124 | |
emilmont | 1:fdd22bb7aa52 | 125 | /* Typecast q15_t pointer to q31_t pointer for coefficient reading in q31_t */ |
emilmont | 1:fdd22bb7aa52 | 126 | pb = pCoeffs; |
emilmont | 1:fdd22bb7aa52 | 127 | |
emilmont | 1:fdd22bb7aa52 | 128 | /* Read the first two samples from the state buffer: x[n-N], x[n-N-1] */ |
emilmont | 1:fdd22bb7aa52 | 129 | x0 = *__SIMD32(px)++; |
emilmont | 1:fdd22bb7aa52 | 130 | |
emilmont | 1:fdd22bb7aa52 | 131 | /* Read the third and forth samples from the state buffer: x[n-N-2], x[n-N-3] */ |
emilmont | 1:fdd22bb7aa52 | 132 | x2 = *__SIMD32(px)++; |
emilmont | 1:fdd22bb7aa52 | 133 | |
emilmont | 1:fdd22bb7aa52 | 134 | /* Loop over the number of taps. Unroll by a factor of 4. |
emilmont | 1:fdd22bb7aa52 | 135 | ** Repeat until we've computed numTaps-(numTaps%4) coefficients. */ |
emilmont | 1:fdd22bb7aa52 | 136 | tapCnt = numTaps >> 2; |
emilmont | 1:fdd22bb7aa52 | 137 | |
emilmont | 1:fdd22bb7aa52 | 138 | while(tapCnt > 0) |
emilmont | 1:fdd22bb7aa52 | 139 | { |
emilmont | 1:fdd22bb7aa52 | 140 | /* Read the first two coefficients using SIMD: b[N] and b[N-1] coefficients */ |
emilmont | 1:fdd22bb7aa52 | 141 | c0 = *__SIMD32(pb)++; |
emilmont | 1:fdd22bb7aa52 | 142 | |
emilmont | 1:fdd22bb7aa52 | 143 | /* acc0 += b[N] * x[n-N] + b[N-1] * x[n-N-1] */ |
emilmont | 1:fdd22bb7aa52 | 144 | acc0 = __SMLAD(x0, c0, acc0); |
emilmont | 1:fdd22bb7aa52 | 145 | |
emilmont | 1:fdd22bb7aa52 | 146 | /* acc2 += b[N] * x[n-N-2] + b[N-1] * x[n-N-3] */ |
emilmont | 1:fdd22bb7aa52 | 147 | acc2 = __SMLAD(x2, c0, acc2); |
emilmont | 1:fdd22bb7aa52 | 148 | |
emilmont | 1:fdd22bb7aa52 | 149 | /* pack x[n-N-1] and x[n-N-2] */ |
emilmont | 1:fdd22bb7aa52 | 150 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 151 | x1 = __PKHBT(x2, x0, 0); |
emilmont | 1:fdd22bb7aa52 | 152 | #else |
emilmont | 1:fdd22bb7aa52 | 153 | x1 = __PKHBT(x0, x2, 0); |
emilmont | 1:fdd22bb7aa52 | 154 | #endif |
emilmont | 1:fdd22bb7aa52 | 155 | |
emilmont | 1:fdd22bb7aa52 | 156 | /* Read state x[n-N-4], x[n-N-5] */ |
emilmont | 1:fdd22bb7aa52 | 157 | x0 = _SIMD32_OFFSET(px); |
emilmont | 1:fdd22bb7aa52 | 158 | |
emilmont | 1:fdd22bb7aa52 | 159 | /* acc1 += b[N] * x[n-N-1] + b[N-1] * x[n-N-2] */ |
emilmont | 1:fdd22bb7aa52 | 160 | acc1 = __SMLADX(x1, c0, acc1); |
emilmont | 1:fdd22bb7aa52 | 161 | |
emilmont | 1:fdd22bb7aa52 | 162 | /* pack x[n-N-3] and x[n-N-4] */ |
emilmont | 1:fdd22bb7aa52 | 163 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 164 | x1 = __PKHBT(x0, x2, 0); |
emilmont | 1:fdd22bb7aa52 | 165 | #else |
emilmont | 1:fdd22bb7aa52 | 166 | x1 = __PKHBT(x2, x0, 0); |
emilmont | 1:fdd22bb7aa52 | 167 | #endif |
emilmont | 1:fdd22bb7aa52 | 168 | |
emilmont | 1:fdd22bb7aa52 | 169 | /* acc3 += b[N] * x[n-N-3] + b[N-1] * x[n-N-4] */ |
emilmont | 1:fdd22bb7aa52 | 170 | acc3 = __SMLADX(x1, c0, acc3); |
emilmont | 1:fdd22bb7aa52 | 171 | |
emilmont | 1:fdd22bb7aa52 | 172 | /* Read coefficients b[N-2], b[N-3] */ |
emilmont | 1:fdd22bb7aa52 | 173 | c0 = *__SIMD32(pb)++; |
emilmont | 1:fdd22bb7aa52 | 174 | |
emilmont | 1:fdd22bb7aa52 | 175 | /* acc0 += b[N-2] * x[n-N-2] + b[N-3] * x[n-N-3] */ |
emilmont | 1:fdd22bb7aa52 | 176 | acc0 = __SMLAD(x2, c0, acc0); |
emilmont | 1:fdd22bb7aa52 | 177 | |
emilmont | 1:fdd22bb7aa52 | 178 | /* Read state x[n-N-6], x[n-N-7] with offset */ |
emilmont | 1:fdd22bb7aa52 | 179 | x2 = _SIMD32_OFFSET(px + 2u); |
emilmont | 1:fdd22bb7aa52 | 180 | |
emilmont | 1:fdd22bb7aa52 | 181 | /* acc2 += b[N-2] * x[n-N-4] + b[N-3] * x[n-N-5] */ |
emilmont | 1:fdd22bb7aa52 | 182 | acc2 = __SMLAD(x0, c0, acc2); |
emilmont | 1:fdd22bb7aa52 | 183 | |
emilmont | 1:fdd22bb7aa52 | 184 | /* acc1 += b[N-2] * x[n-N-3] + b[N-3] * x[n-N-4] */ |
emilmont | 1:fdd22bb7aa52 | 185 | acc1 = __SMLADX(x1, c0, acc1); |
emilmont | 1:fdd22bb7aa52 | 186 | |
emilmont | 1:fdd22bb7aa52 | 187 | /* pack x[n-N-5] and x[n-N-6] */ |
emilmont | 1:fdd22bb7aa52 | 188 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 189 | x1 = __PKHBT(x2, x0, 0); |
emilmont | 1:fdd22bb7aa52 | 190 | #else |
emilmont | 1:fdd22bb7aa52 | 191 | x1 = __PKHBT(x0, x2, 0); |
emilmont | 1:fdd22bb7aa52 | 192 | #endif |
emilmont | 1:fdd22bb7aa52 | 193 | |
emilmont | 1:fdd22bb7aa52 | 194 | /* acc3 += b[N-2] * x[n-N-5] + b[N-3] * x[n-N-6] */ |
emilmont | 1:fdd22bb7aa52 | 195 | acc3 = __SMLADX(x1, c0, acc3); |
emilmont | 1:fdd22bb7aa52 | 196 | |
emilmont | 1:fdd22bb7aa52 | 197 | /* Update state pointer for next state reading */ |
emilmont | 1:fdd22bb7aa52 | 198 | px += 4u; |
emilmont | 1:fdd22bb7aa52 | 199 | |
emilmont | 1:fdd22bb7aa52 | 200 | /* Decrement tap count */ |
emilmont | 1:fdd22bb7aa52 | 201 | tapCnt--; |
emilmont | 1:fdd22bb7aa52 | 202 | |
emilmont | 1:fdd22bb7aa52 | 203 | } |
emilmont | 1:fdd22bb7aa52 | 204 | |
emilmont | 1:fdd22bb7aa52 | 205 | /* If the filter length is not a multiple of 4, compute the remaining filter taps. |
emilmont | 1:fdd22bb7aa52 | 206 | ** This is always be 2 taps since the filter length is even. */ |
emilmont | 1:fdd22bb7aa52 | 207 | if((numTaps & 0x3u) != 0u) |
emilmont | 1:fdd22bb7aa52 | 208 | { |
emilmont | 1:fdd22bb7aa52 | 209 | |
emilmont | 1:fdd22bb7aa52 | 210 | /* Read last two coefficients */ |
emilmont | 1:fdd22bb7aa52 | 211 | c0 = *__SIMD32(pb)++; |
emilmont | 1:fdd22bb7aa52 | 212 | |
emilmont | 1:fdd22bb7aa52 | 213 | /* Perform the multiply-accumulates */ |
emilmont | 1:fdd22bb7aa52 | 214 | acc0 = __SMLAD(x0, c0, acc0); |
emilmont | 1:fdd22bb7aa52 | 215 | acc2 = __SMLAD(x2, c0, acc2); |
emilmont | 1:fdd22bb7aa52 | 216 | |
emilmont | 1:fdd22bb7aa52 | 217 | /* pack state variables */ |
emilmont | 1:fdd22bb7aa52 | 218 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 219 | x1 = __PKHBT(x2, x0, 0); |
emilmont | 1:fdd22bb7aa52 | 220 | #else |
emilmont | 1:fdd22bb7aa52 | 221 | x1 = __PKHBT(x0, x2, 0); |
emilmont | 1:fdd22bb7aa52 | 222 | #endif |
emilmont | 1:fdd22bb7aa52 | 223 | |
emilmont | 1:fdd22bb7aa52 | 224 | /* Read last state variables */ |
emilmont | 1:fdd22bb7aa52 | 225 | x0 = *__SIMD32(px); |
emilmont | 1:fdd22bb7aa52 | 226 | |
emilmont | 1:fdd22bb7aa52 | 227 | /* Perform the multiply-accumulates */ |
emilmont | 1:fdd22bb7aa52 | 228 | acc1 = __SMLADX(x1, c0, acc1); |
emilmont | 1:fdd22bb7aa52 | 229 | |
emilmont | 1:fdd22bb7aa52 | 230 | /* pack state variables */ |
emilmont | 1:fdd22bb7aa52 | 231 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 232 | x1 = __PKHBT(x0, x2, 0); |
emilmont | 1:fdd22bb7aa52 | 233 | #else |
emilmont | 1:fdd22bb7aa52 | 234 | x1 = __PKHBT(x2, x0, 0); |
emilmont | 1:fdd22bb7aa52 | 235 | #endif |
emilmont | 1:fdd22bb7aa52 | 236 | |
emilmont | 1:fdd22bb7aa52 | 237 | /* Perform the multiply-accumulates */ |
emilmont | 1:fdd22bb7aa52 | 238 | acc3 = __SMLADX(x1, c0, acc3); |
emilmont | 1:fdd22bb7aa52 | 239 | } |
emilmont | 1:fdd22bb7aa52 | 240 | |
emilmont | 1:fdd22bb7aa52 | 241 | /* The results in the 4 accumulators are in 2.30 format. Convert to 1.15 with saturation. |
emilmont | 1:fdd22bb7aa52 | 242 | ** Then store the 4 outputs in the destination buffer. */ |
emilmont | 1:fdd22bb7aa52 | 243 | |
emilmont | 1:fdd22bb7aa52 | 244 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 245 | |
emilmont | 1:fdd22bb7aa52 | 246 | *__SIMD32(pDst)++ = |
emilmont | 1:fdd22bb7aa52 | 247 | __PKHBT(__SSAT((acc0 >> 15), 16), __SSAT((acc1 >> 15), 16), 16); |
emilmont | 1:fdd22bb7aa52 | 248 | |
emilmont | 1:fdd22bb7aa52 | 249 | *__SIMD32(pDst)++ = |
emilmont | 1:fdd22bb7aa52 | 250 | __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16); |
emilmont | 1:fdd22bb7aa52 | 251 | |
emilmont | 1:fdd22bb7aa52 | 252 | #else |
emilmont | 1:fdd22bb7aa52 | 253 | |
emilmont | 1:fdd22bb7aa52 | 254 | *__SIMD32(pDst)++ = |
emilmont | 1:fdd22bb7aa52 | 255 | __PKHBT(__SSAT((acc1 >> 15), 16), __SSAT((acc0 >> 15), 16), 16); |
emilmont | 1:fdd22bb7aa52 | 256 | |
emilmont | 1:fdd22bb7aa52 | 257 | *__SIMD32(pDst)++ = |
emilmont | 1:fdd22bb7aa52 | 258 | __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16); |
emilmont | 1:fdd22bb7aa52 | 259 | |
emilmont | 1:fdd22bb7aa52 | 260 | |
emilmont | 1:fdd22bb7aa52 | 261 | #endif /* #ifndef ARM_MATH_BIG_ENDIAN */ |
emilmont | 1:fdd22bb7aa52 | 262 | |
emilmont | 1:fdd22bb7aa52 | 263 | /* Advance the state pointer by 4 to process the next group of 4 samples */ |
emilmont | 1:fdd22bb7aa52 | 264 | pState = pState + 4u; |
emilmont | 1:fdd22bb7aa52 | 265 | |
emilmont | 1:fdd22bb7aa52 | 266 | /* Decrement the loop counter */ |
emilmont | 1:fdd22bb7aa52 | 267 | blkCnt--; |
emilmont | 1:fdd22bb7aa52 | 268 | } |
emilmont | 1:fdd22bb7aa52 | 269 | |
emilmont | 1:fdd22bb7aa52 | 270 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here. |
emilmont | 1:fdd22bb7aa52 | 271 | ** No loop unrolling is used. */ |
emilmont | 1:fdd22bb7aa52 | 272 | blkCnt = blockSize % 0x4u; |
emilmont | 1:fdd22bb7aa52 | 273 | while(blkCnt > 0u) |
emilmont | 1:fdd22bb7aa52 | 274 | { |
emilmont | 1:fdd22bb7aa52 | 275 | /* Copy two samples into state buffer */ |
emilmont | 1:fdd22bb7aa52 | 276 | *pStateCurnt++ = *pSrc++; |
emilmont | 1:fdd22bb7aa52 | 277 | |
emilmont | 1:fdd22bb7aa52 | 278 | /* Set the accumulator to zero */ |
emilmont | 1:fdd22bb7aa52 | 279 | acc0 = 0; |
emilmont | 1:fdd22bb7aa52 | 280 | |
emilmont | 1:fdd22bb7aa52 | 281 | /* Use SIMD to hold states and coefficients */ |
emilmont | 1:fdd22bb7aa52 | 282 | px = pState; |
emilmont | 1:fdd22bb7aa52 | 283 | pb = pCoeffs; |
emilmont | 1:fdd22bb7aa52 | 284 | |
emilmont | 1:fdd22bb7aa52 | 285 | tapCnt = numTaps >> 1u; |
emilmont | 1:fdd22bb7aa52 | 286 | |
emilmont | 1:fdd22bb7aa52 | 287 | do |
emilmont | 1:fdd22bb7aa52 | 288 | { |
emilmont | 1:fdd22bb7aa52 | 289 | |
emilmont | 1:fdd22bb7aa52 | 290 | acc0 += (q31_t) * px++ * *pb++; |
emilmont | 2:da51fb522205 | 291 | acc0 += (q31_t) * px++ * *pb++; |
emilmont | 1:fdd22bb7aa52 | 292 | |
emilmont | 1:fdd22bb7aa52 | 293 | tapCnt--; |
emilmont | 1:fdd22bb7aa52 | 294 | } |
emilmont | 1:fdd22bb7aa52 | 295 | while(tapCnt > 0u); |
emilmont | 1:fdd22bb7aa52 | 296 | |
emilmont | 1:fdd22bb7aa52 | 297 | /* The result is in 2.30 format. Convert to 1.15 with saturation. |
emilmont | 1:fdd22bb7aa52 | 298 | ** Then store the output in the destination buffer. */ |
emilmont | 1:fdd22bb7aa52 | 299 | *pDst++ = (q15_t) (__SSAT((acc0 >> 15), 16)); |
emilmont | 1:fdd22bb7aa52 | 300 | |
emilmont | 1:fdd22bb7aa52 | 301 | /* Advance state pointer by 1 for the next sample */ |
emilmont | 1:fdd22bb7aa52 | 302 | pState = pState + 1u; |
emilmont | 1:fdd22bb7aa52 | 303 | |
emilmont | 1:fdd22bb7aa52 | 304 | /* Decrement the loop counter */ |
emilmont | 1:fdd22bb7aa52 | 305 | blkCnt--; |
emilmont | 1:fdd22bb7aa52 | 306 | } |
emilmont | 1:fdd22bb7aa52 | 307 | |
emilmont | 1:fdd22bb7aa52 | 308 | /* Processing is complete. |
emilmont | 1:fdd22bb7aa52 | 309 | ** Now copy the last numTaps - 1 samples to the satrt of the state buffer. |
emilmont | 1:fdd22bb7aa52 | 310 | ** This prepares the state buffer for the next function call. */ |
emilmont | 1:fdd22bb7aa52 | 311 | |
emilmont | 1:fdd22bb7aa52 | 312 | /* Points to the start of the state buffer */ |
emilmont | 1:fdd22bb7aa52 | 313 | pStateCurnt = S->pState; |
emilmont | 1:fdd22bb7aa52 | 314 | |
emilmont | 1:fdd22bb7aa52 | 315 | /* Calculation of count for copying integer writes */ |
emilmont | 1:fdd22bb7aa52 | 316 | tapCnt = (numTaps - 1u) >> 2; |
emilmont | 1:fdd22bb7aa52 | 317 | |
emilmont | 1:fdd22bb7aa52 | 318 | while(tapCnt > 0u) |
emilmont | 1:fdd22bb7aa52 | 319 | { |
emilmont | 1:fdd22bb7aa52 | 320 | *pStateCurnt++ = *pState++; |
emilmont | 1:fdd22bb7aa52 | 321 | *pStateCurnt++ = *pState++; |
emilmont | 1:fdd22bb7aa52 | 322 | *pStateCurnt++ = *pState++; |
emilmont | 1:fdd22bb7aa52 | 323 | *pStateCurnt++ = *pState++; |
emilmont | 1:fdd22bb7aa52 | 324 | |
emilmont | 1:fdd22bb7aa52 | 325 | tapCnt--; |
emilmont | 1:fdd22bb7aa52 | 326 | |
emilmont | 1:fdd22bb7aa52 | 327 | } |
emilmont | 1:fdd22bb7aa52 | 328 | |
emilmont | 1:fdd22bb7aa52 | 329 | /* Calculation of count for remaining q15_t data */ |
emilmont | 1:fdd22bb7aa52 | 330 | tapCnt = (numTaps - 1u) % 0x4u; |
emilmont | 1:fdd22bb7aa52 | 331 | |
emilmont | 1:fdd22bb7aa52 | 332 | /* copy remaining data */ |
emilmont | 1:fdd22bb7aa52 | 333 | while(tapCnt > 0u) |
emilmont | 1:fdd22bb7aa52 | 334 | { |
emilmont | 1:fdd22bb7aa52 | 335 | *pStateCurnt++ = *pState++; |
emilmont | 1:fdd22bb7aa52 | 336 | |
emilmont | 1:fdd22bb7aa52 | 337 | /* Decrement the loop counter */ |
emilmont | 1:fdd22bb7aa52 | 338 | tapCnt--; |
emilmont | 1:fdd22bb7aa52 | 339 | } |
emilmont | 1:fdd22bb7aa52 | 340 | |
emilmont | 1:fdd22bb7aa52 | 341 | } |
emilmont | 1:fdd22bb7aa52 | 342 | |
emilmont | 1:fdd22bb7aa52 | 343 | /** |
emilmont | 1:fdd22bb7aa52 | 344 | * @} end of FIR group |
emilmont | 1:fdd22bb7aa52 | 345 | */ |