k60-keil

K60-Keil版本（下载安装MDK4.23）

      gnext2 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + gnext1;
      gnext2 = __SSAT(gnext2, 16);
      gnext1 = (q31_t) ((fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
      gnext1 = __SSAT(gnext1, 16);


      /* Read g2(n-1), g4(n-1) .... from state */
      gcurnt1 = *px;

      /* save g4(n) in state buffer */
      *px++ = (q15_t) gnext4;

      /* Sample processing for K5, K9.... */
      /* Process first sample for 5th, 9th .. tap */
      /* f5(n) = f4(n) +  K5 * g4(n-1) */
      fcurnt1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fnext1;
      fcurnt1 = __SSAT(fcurnt1, 16);

      /* Process second sample for 5th, 9th .. tap */
      fcurnt2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fnext2;
      fcurnt2 = __SSAT(fcurnt2, 16);

      /* Process third sample for 5th, 9th .. tap */
      fcurnt3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fnext3;
      fcurnt3 = __SSAT(fcurnt3, 16);

      /* Process fourth sample for 5th, 9th .. tap */
      fcurnt4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fnext4;
      fcurnt4 = __SSAT(fcurnt4, 16);

      /* Calculation of state values for next stage */
      /* g5(n) = f4(n) * K5  +  g4(n-1) */
      gnext4 = (q31_t) ((fnext4 * (*pk)) >> 15u) + gnext3;
      gnext4 = __SSAT(gnext4, 16);
      gnext3 = (q31_t) ((fnext3 * (*pk)) >> 15u) + gnext2;
      gnext3 = __SSAT(gnext3, 16);
      gnext2 = (q31_t) ((fnext2 * (*pk)) >> 15u) + gnext1;
      gnext2 = __SSAT(gnext2, 16);
      gnext1 = (q31_t) ((fnext1 * (*pk++)) >> 15u) + gcurnt1;
      gnext1 = __SSAT(gnext1, 16);

      stageCnt--;
    }

    /* If the (filter length -1) is not a multiple of 4, compute the remaining filter taps */
    stageCnt = (numStages - 1u) % 0x4u;

    while(stageCnt > 0u)
    {
      gcurnt1 = *px;

      /* save g value in state buffer */
      *px++ = (q15_t) gnext4;

      /* Process four samples for last three taps here */
      fnext1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fcurnt1;
      fnext1 = __SSAT(fnext1, 16);
      fnext2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fcurnt2;
      fnext2 = __SSAT(fnext2, 16);

      fnext3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fcurnt3;
      fnext3 = __SSAT(fnext3, 16);

      fnext4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fcurnt4;
      fnext4 = __SSAT(fnext4, 16);

      /* g1(n) = f0(n) * K1  +  g0(n-1) */
      gnext4 = (q31_t) ((fcurnt4 * (*pk)) >> 15u) + gnext3;
      gnext4 = __SSAT(gnext4, 16);
      gnext3 = (q31_t) ((fcurnt3 * (*pk)) >> 15u) + gnext2;
      gnext3 = __SSAT(gnext3, 16);
      gnext2 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + gnext1;
      gnext2 = __SSAT(gnext2, 16);
      gnext1 = (q31_t) ((fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
      gnext1 = __SSAT(gnext1, 16);

      /* Update of f values for next coefficient set processing */
      fcurnt1 = fnext1;
      fcurnt2 = fnext2;
      fcurnt3 = fnext3;
      fcurnt4 = fnext4;

      stageCnt--;

    }

    /* The results in the 4 accumulators, store in the destination buffer. */
    /* y(n) = fN(n) */

#ifndef  ARM_MATH_BIG_ENDIAN

    *__SIMD32(pDst)++ = __PKHBT(fcurnt1, fcurnt2, 16);
    *__SIMD32(pDst)++ = __PKHBT(fcurnt3, fcurnt4, 16);

#else

    *__SIMD32(pDst)++ = __PKHBT(fcurnt2, fcurnt1, 16);
    *__SIMD32(pDst)++ = __PKHBT(fcurnt4, fcurnt3, 16);

#endif /*      #ifndef  ARM_MATH_BIG_ENDIAN    */

    blkCnt--;
  }

  /* If the blockSize is not a multiple of 4, compute any remaining output samples here.   
   ** No loop unrolling is used. */
  blkCnt = blockSize % 0x4u;

  while(blkCnt > 0u)
  {
    /* f0(n) = x(n) */
    fcurnt1 = *pSrc++;

    /* Initialize coeff pointer */
    pk = (pCoeffs);

    /* Initialize state pointer */
    px = pState;

    /* read g2(n) from state buffer */
    gcurnt1 = *px;

    /* for sample 1 processing */
    /* f1(n) = f0(n) +  K1 * g0(n-1) */
    fnext1 = (((q31_t) gcurnt1 * (*pk)) >> 15u) + fcurnt1;
    fnext1 = __SSAT(fnext1, 16);


    /* g1(n) = f0(n) * K1  +  g0(n-1) */
    gnext1 = (((q31_t) fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
    gnext1 = __SSAT(gnext1, 16);

    /* save g1(n) in state buffer */
    *px++ = (q15_t) fcurnt1;

    /* f1(n) is saved in fcurnt1   
       for next stage processing */
    fcurnt1 = fnext1;

    stageCnt = (numStages - 1u);

    /* stage loop */
    while(stageCnt > 0u)
    {
      /* read g2(n) from state buffer */
      gcurnt1 = *px;

      /* save g1(n) in state buffer */
      *px++ = (q15_t) gnext1;

      /* Sample processing for K2, K3.... */
      /* f2(n) = f1(n) +  K2 * g1(n-1) */
      fnext1 = (((q31_t) gcurnt1 * (*pk)) >> 15u) + fcurnt1;
      fnext1 = __SSAT(fnext1, 16);

      /* g2(n) = f1(n) * K2  +  g1(n-1) */
      gnext1 = (((q31_t) fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
      gnext1 = __SSAT(gnext1, 16);


      /* f1(n) is saved in fcurnt1   
         for next stage processing */
      fcurnt1 = fnext1;

      stageCnt--;

    }

    /* y(n) = fN(n) */
    *pDst++ = __SSAT(fcurnt1, 16);


    blkCnt--;

  }

#else

  /* Run the below code for Cortex-M0 */

  q31_t fcurnt, fnext, gcurnt, gnext;            /* temporary variables */
  uint32_t numStages = S->numStages;             /* Length of the filter */
  uint32_t blkCnt, stageCnt;                     /* temporary variables for counts */

  pState = &S->pState[0];

  blkCnt = blockSize;

  while(blkCnt > 0u)
  {
    /* f0(n) = x(n) */
    fcurnt = *pSrc++;

    /* Initialize coeff pointer */
    pk = (pCoeffs);

    /* Initialize state pointer */
    px = pState;

    /* read g0(n-1) from state buffer */
    gcurnt = *px;

    /* for sample 1 processing */
    /* f1(n) = f0(n) +  K1 * g0(n-1) */
    fnext = ((gcurnt * (*pk)) >> 15u) + fcurnt;
    fnext = __SSAT(fnext, 16);


    /* g1(n) = f0(n) * K1  +  g0(n-1) */
    gnext = ((fcurnt * (*pk++)) >> 15u) + gcurnt;
    gnext = __SSAT(gnext, 16);

    /* save f0(n) in state buffer */
    *px++ = (q15_t) fcurnt;

    /* f1(n) is saved in fcurnt           
       for next stage processing */
    fcurnt = fnext;

    stageCnt = (numStages - 1u);

    /* stage loop */
    while(stageCnt > 0u)
    {
      /* read g1(n-1) from state buffer */
      gcurnt = *px;

      /* save g0(n-1) in state buffer */
      *px++ = (q15_t) gnext;

      /* Sample processing for K2, K3.... */
      /* f2(n) = f1(n) +  K2 * g1(n-1) */
      fnext = ((gcurnt * (*pk)) >> 15u) + fcurnt;
      fnext = __SSAT(fnext, 16);

      /* g2(n) = f1(n) * K2  +  g1(n-1) */
      gnext = ((fcurnt * (*pk++)) >> 15u) + gcurnt;
      gnext = __SSAT(gnext, 16);


      /* f1(n) is saved in fcurnt           
         for next stage processing */
      fcurnt = fnext;

      stageCnt--;

    }

    /* y(n) = fN(n) */
    *pDst++ = __SSAT(fcurnt, 16);


    blkCnt--;

  }

#endif /*   #ifndef ARM_MATH_CM0 */

}

/**   
 * @} end of FIR_Lattice group   
 */