00001 /* ---------------------------------------------------------------------- 00002 * Copyright (C) 2010 ARM Limited. All rights reserved. 00003 * 00004 * $Date: 15. July 2011 00005 * $Revision: V1.0.10 00006 * 00007 * Project: CMSIS DSP Library 00008 * Title: arm_mat_scale_q31.c 00009 * 00010 * Description: Multiplies a Q31 matrix by a scalar. 00011 * 00012 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0 00013 * 00014 * Version 1.0.10 2011/7/15 00015 * Big Endian support added and Merged M0 and M3/M4 Source code. 00016 * 00017 * Version 1.0.3 2010/11/29 00018 * Re-organized the CMSIS folders and updated documentation. 00019 * 00020 * Version 1.0.2 2010/11/11 00021 * Documentation updated. 00022 * 00023 * Version 1.0.1 2010/10/05 00024 * Production release and review comments incorporated. 00025 * 00026 * Version 1.0.0 2010/09/20 00027 * Production release and review comments incorporated. 00028 * 00029 * Version 0.0.5 2010/04/26 00030 * incorporated review comments and updated with latest CMSIS layer 00031 * 00032 * Version 0.0.3 2010/03/10 00033 * Initial version 00034 * -------------------------------------------------------------------- */ 00035 00036 #include "arm_math.h" 00037 00063 arm_status arm_mat_scale_q31( 00064 const arm_matrix_instance_q31 * pSrc, 00065 q31_t scaleFract, 00066 int32_t shift, 00067 arm_matrix_instance_q31 * pDst) 00068 { 00069 q31_t *pIn = pSrc->pData; /* input data matrix pointer */ 00070 q31_t *pOut = pDst->pData; /* output data matrix pointer */ 00071 q63_t out; /* temporary variable to hold output value */ 00072 uint32_t numSamples; /* total number of elements in the matrix */ 00073 int32_t totShift = 31 - shift; /* shift to apply after scaling */ 00074 uint32_t blkCnt; /* loop counters */ 00075 arm_status status; /* status of matrix scaling */ 00076 00077 #ifdef ARM_MATH_MATRIX_CHECK 00078 00079 00080 /* Check for matrix mismatch */ 00081 if((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols)) 00082 { 00083 /* Set status as ARM_MATH_SIZE_MISMATCH */ 00084 status = ARM_MATH_SIZE_MISMATCH; 00085 } 00086 else 00087 #endif /* #ifdef ARM_MATH_MATRIX_CHECK */ 00088 00089 { 00090 /* Total number of samples in the input matrix */ 00091 numSamples = (uint32_t) pSrc->numRows * pSrc->numCols; 00092 00093 #ifndef ARM_MATH_CM0 00094 00095 /* Run the below code for Cortex-M4 and Cortex-M3 */ 00096 00097 /* Loop Unrolling */ 00098 blkCnt = numSamples >> 2u; 00099 00100 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00101 ** a second loop below computes the remaining 1 to 3 samples. */ 00102 while(blkCnt > 0u) 00103 { 00104 /* C(m,n) = A(m,n) * k */ 00105 /* Scale, saturate and then store the results in the destination buffer. */ 00106 out = ((q63_t) * pIn++ * scaleFract) >> totShift; 00107 *pOut++ = clip_q63_to_q31(out); 00108 out = ((q63_t) * pIn++ * scaleFract) >> totShift; 00109 *pOut++ = clip_q63_to_q31(out); 00110 out = ((q63_t) * pIn++ * scaleFract) >> totShift; 00111 *pOut++ = clip_q63_to_q31(out); 00112 out = ((q63_t) * pIn++ * scaleFract) >> totShift; 00113 *pOut++ = clip_q63_to_q31(out); 00114 00115 /* Decrement the numSamples loop counter */ 00116 blkCnt--; 00117 } 00118 00119 /* If the numSamples is not a multiple of 4, compute any remaining output samples here. 00120 ** No loop unrolling is used. */ 00121 blkCnt = numSamples % 0x4u; 00122 00123 #else 00124 00125 /* Run the below code for Cortex-M0 */ 00126 00127 /* Initialize blkCnt with number of samples */ 00128 blkCnt = numSamples; 00129 00130 #endif /* #ifndef ARM_MATH_CM0 */ 00131 00132 while(blkCnt > 0u) 00133 { 00134 /* C(m,n) = A(m,n) * k */ 00135 /* Scale, saturate and then store the results in the destination buffer. */ 00136 out = ((q63_t) * pIn++ * scaleFract) >> totShift; 00137 *pOut++ = clip_q63_to_q31(out); 00138 00139 /* Decrement the numSamples loop counter */ 00140 blkCnt--; 00141 } 00142 /* Set status as ARM_MATH_SUCCESS */ 00143 status = ARM_MATH_SUCCESS; 00144 } 00145 00146 /* Return to application */ 00147 return (status); 00148 } 00149