arm_scale_q7.c

00001 /* ----------------------------------------------------------------------
00002  * Project:      CMSIS DSP Library
00003  * Title:        arm_scale_q7.c
00004  * Description:  Multiplies a Q7 vector by a scalar
00005  *
00006  * $Date:        27. January 2017
00007  * $Revision:    V.1.5.1
00008  *
00009  * Target Processor: Cortex-M cores
00010  * -------------------------------------------------------------------- */
00011 /*
00012  * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved.
00013  *
00014  * SPDX-License-Identifier: Apache-2.0
00015  *
00016  * Licensed under the Apache License, Version 2.0 (the License); you may
00017  * not use this file except in compliance with the License.
00018  * You may obtain a copy of the License at
00019  *
00020  * www.apache.org/licenses/LICENSE-2.0
00021  *
00022  * Unless required by applicable law or agreed to in writing, software
00023  * distributed under the License is distributed on an AS IS BASIS, WITHOUT
00024  * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
00025  * See the License for the specific language governing permissions and
00026  * limitations under the License.
00027  */
00028 
00029 #include "arm_math.h"
00030 
00031 /**
00032  * @ingroup groupMath
00033  */
00034 
00035 /**
00036  * @addtogroup scale
00037  * @{
00038  */
00039 
00040 /**
00041  * @brief Multiplies a Q7 vector by a scalar.
00042  * @param[in]       *pSrc points to the input vector
00043  * @param[in]       scaleFract fractional portion of the scale value
00044  * @param[in]       shift number of bits to shift the result by
00045  * @param[out]      *pDst points to the output vector
00046  * @param[in]       blockSize number of samples in the vector
00047  * @return none.
00048  *
00049  * <b>Scaling and Overflow Behavior:</b>
00050  * \par
00051  * The input data <code>*pSrc</code> and <code>scaleFract</code> are in 1.7 format.
00052  * These are multiplied to yield a 2.14 intermediate result and this is shifted with saturation to 1.7 format.
00053  */
00054 
00055 void arm_scale_q7(
00056   q7_t * pSrc,
00057   q7_t scaleFract,
00058   int8_t shift,
00059   q7_t * pDst,
00060   uint32_t blockSize)
00061 {
00062   int8_t kShift = 7 - shift;                     /* shift to apply after scaling */
00063   uint32_t blkCnt;                               /* loop counter */
00064 
00065 #if defined (ARM_MATH_DSP)
00066 
00067 /* Run the below code for Cortex-M4 and Cortex-M3 */
00068   q7_t in1, in2, in3, in4, out1, out2, out3, out4;      /* Temporary variables to store input & output */
00069 
00070 
00071   /*loop Unrolling */
00072   blkCnt = blockSize >> 2U;
00073 
00074 
00075   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
00076    ** a second loop below computes the remaining 1 to 3 samples. */
00077   while (blkCnt > 0U)
00078   {
00079     /* Reading 4 inputs from memory */
00080     in1 = *pSrc++;
00081     in2 = *pSrc++;
00082     in3 = *pSrc++;
00083     in4 = *pSrc++;
00084 
00085     /* C = A * scale */
00086     /* Scale the inputs and then store the results in the temporary variables. */
00087     out1 = (q7_t) (__SSAT(((in1) * scaleFract) >> kShift, 8));
00088     out2 = (q7_t) (__SSAT(((in2) * scaleFract) >> kShift, 8));
00089     out3 = (q7_t) (__SSAT(((in3) * scaleFract) >> kShift, 8));
00090     out4 = (q7_t) (__SSAT(((in4) * scaleFract) >> kShift, 8));
00091 
00092     /* Packing the individual outputs into 32bit and storing in
00093      * destination buffer in single write */
00094     *__SIMD32(pDst)++ = __PACKq7(out1, out2, out3, out4);
00095 
00096     /* Decrement the loop counter */
00097     blkCnt--;
00098   }
00099 
00100   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
00101    ** No loop unrolling is used. */
00102   blkCnt = blockSize % 0x4U;
00103 
00104   while (blkCnt > 0U)
00105   {
00106     /* C = A * scale */
00107     /* Scale the input and then store the result in the destination buffer. */
00108     *pDst++ = (q7_t) (__SSAT(((*pSrc++) * scaleFract) >> kShift, 8));
00109 
00110     /* Decrement the loop counter */
00111     blkCnt--;
00112   }
00113 
00114 #else
00115 
00116   /* Run the below code for Cortex-M0 */
00117 
00118   /* Initialize blkCnt with number of samples */
00119   blkCnt = blockSize;
00120 
00121   while (blkCnt > 0U)
00122   {
00123     /* C = A * scale */
00124     /* Scale the input and then store the result in the destination buffer. */
00125     *pDst++ = (q7_t) (__SSAT((((q15_t) * pSrc++ * scaleFract) >> kShift), 8));
00126 
00127     /* Decrement the loop counter */
00128     blkCnt--;
00129   }
00130 
00131 #endif /* #if defined (ARM_MATH_DSP) */
00132 
00133 }
00134 
00135 /**
00136  * @} end of scale group
00137  */
00138