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_q15.c 00009 * 00010 * Description: Multiplies a Q15 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_q15( 00064 const arm_matrix_instance_q15 * pSrc, 00065 q15_t scaleFract, 00066 int32_t shift, 00067 arm_matrix_instance_q15 * pDst) 00068 { 00069 q15_t *pIn = pSrc->pData; /* input data matrix pointer */ 00070 q15_t *pOut = pDst->pData; /* output data matrix pointer */ 00071 uint32_t numSamples; /* total number of elements in the matrix */ 00072 int32_t totShift = 15 - shift; /* total shift to apply after scaling */ 00073 uint32_t blkCnt; /* loop counters */ 00074 arm_status status; /* status of matrix scaling */ 00075 00076 #ifdef ARM_MATH_MATRIX_CHECK 00077 00078 00079 /* Check for matrix mismatch */ 00080 if((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols)) 00081 { 00082 /* Set status as ARM_MATH_SIZE_MISMATCH */ 00083 status = ARM_MATH_SIZE_MISMATCH; 00084 } 00085 else 00086 #endif /* #ifdef ARM_MATH_MATRIX_CHECK */ 00087 00088 { 00089 /* Total number of samples in the input matrix */ 00090 numSamples = (uint32_t) pSrc->numRows * pSrc->numCols; 00091 00092 #ifndef ARM_MATH_CM0 00093 00094 /* Run the below code for Cortex-M4 and Cortex-M3 */ 00095 /* Loop Unrolling */ 00096 blkCnt = numSamples >> 2; 00097 00098 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00099 ** a second loop below computes the remaining 1 to 3 samples. */ 00100 while(blkCnt > 0u) 00101 { 00102 /* C(m,n) = A(m,n) * k */ 00103 /* Scale, saturate and then store the results in the destination buffer. */ 00104 *pOut++ = 00105 (q15_t) (__SSAT(((q31_t) (*pIn++) * scaleFract) >> totShift, 16)); 00106 *pOut++ = 00107 (q15_t) (__SSAT(((q31_t) (*pIn++) * scaleFract) >> totShift, 16)); 00108 *pOut++ = 00109 (q15_t) (__SSAT(((q31_t) (*pIn++) * scaleFract) >> totShift, 16)); 00110 *pOut++ = 00111 (q15_t) (__SSAT(((q31_t) (*pIn++) * scaleFract) >> totShift, 16)); 00112 00113 /* Decrement the numSamples loop counter */ 00114 blkCnt--; 00115 } 00116 00117 /* If the numSamples is not a multiple of 4, compute any remaining output samples here. 00118 ** No loop unrolling is used. */ 00119 blkCnt = numSamples % 0x4u; 00120 00121 #else 00122 00123 /* Run the below code for Cortex-M0 */ 00124 00125 /* Initialize blkCnt with number of samples */ 00126 blkCnt = numSamples; 00127 00128 #endif /* #ifndef ARM_MATH_CM0 */ 00129 00130 while(blkCnt > 0u) 00131 { 00132 /* C(m,n) = A(m,n) * k */ 00133 /* Scale, saturate and then store the results in the destination buffer. */ 00134 *pOut++ = 00135 (q15_t) (__SSAT(((q31_t) (*pIn++) * scaleFract) >> totShift, 16)); 00136 00137 /* Decrement the numSamples loop counter */ 00138 blkCnt--; 00139 } 00140 /* Set status as ARM_MATH_SUCCESS */ 00141 status = ARM_MATH_SUCCESS; 00142 } 00143 00144 /* Return to application */ 00145 return (status); 00146 } 00147