Fork of mbed-dsp. CMSIS-DSP library of supporting NEON

Dependents:   mbed-os-example-cmsis_dsp_neon

Fork of mbed-dsp by mbed official

Information

Japanese version is available in lower part of this page.
このページの後半に日本語版が用意されています.

CMSIS-DSP of supporting NEON

What is this ?

A library for CMSIS-DSP of supporting NEON.
We supported the NEON to CMSIS-DSP Ver1.4.3(CMSIS V4.1) that ARM supplied, has achieved the processing speed improvement.
If you use the mbed-dsp library, you can use to replace this library.
CMSIS-DSP of supporting NEON is provied as a library.

Library Creation environment

CMSIS-DSP library of supporting NEON was created by the following environment.

  • Compiler
    ARMCC Version 5.03
  • Compile option switch[C Compiler]
   -DARM_MATH_MATRIX_CHECK -DARM_MATH_ROUNDING -O3 -Otime --cpu=Cortex-A9 --littleend --arm 
   --apcs=/interwork --no_unaligned_access --fpu=vfpv3_fp16 --fpmode=fast --apcs=/hardfp 
   --vectorize --asm
  • Compile option switch[Assembler]
   --cpreproc --cpu=Cortex-A9 --littleend --arm --apcs=/interwork --no_unaligned_access 
   --fpu=vfpv3_fp16 --fpmode=fast --apcs=/hardfp


Effects of NEON support

In the data which passes to each function, large size will be expected more effective than small size.
Also if the data is a multiple of 16, effect will be expected in every function in the CMSIS-DSP.


NEON対応CMSIS-DSP

概要

NEON対応したCMSIS-DSPのライブラリです。
ARM社提供のCMSIS-DSP Ver1.4.3(CMSIS V4.1)をターゲットにNEON対応を行ない、処理速度向上を実現しております。
mbed-dspライブラリを使用している場合は、本ライブラリに置き換えて使用することができます。
NEON対応したCMSIS-DSPはライブラリで提供します。

ライブラリ作成環境

NEON対応CMSIS-DSPライブラリは、以下の環境で作成しています。

  • コンパイラ
    ARMCC Version 5.03
  • コンパイルオプションスイッチ[C Compiler]
   -DARM_MATH_MATRIX_CHECK -DARM_MATH_ROUNDING -O3 -Otime --cpu=Cortex-A9 --littleend --arm 
   --apcs=/interwork --no_unaligned_access --fpu=vfpv3_fp16 --fpmode=fast --apcs=/hardfp 
   --vectorize --asm
  • コンパイルオプションスイッチ[Assembler]
   --cpreproc --cpu=Cortex-A9 --littleend --arm --apcs=/interwork --no_unaligned_access 
   --fpu=vfpv3_fp16 --fpmode=fast --apcs=/hardfp


NEON対応による効果について

CMSIS-DSP内の各関数へ渡すデータは、小さいサイズよりも大きいサイズの方が効果が見込めます。
また、16の倍数のデータであれば、CMSIS-DSP内のどの関数でも効果が見込めます。


Revision:
1:fdd22bb7aa52
Child:
2:da51fb522205
diff -r 83d0537c7d84 -r fdd22bb7aa52 cmsis_dsp/FastMathFunctions/arm_sin_q31.c
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/cmsis_dsp/FastMathFunctions/arm_sin_q31.c	Wed Nov 28 12:30:09 2012 +0000
@@ -0,0 +1,240 @@
+/* ----------------------------------------------------------------------    
+* Copyright (C) 2010 ARM Limited. All rights reserved.    
+*    
+* $Date:        15. February 2012  
+* $Revision:     V1.1.0  
+*    
+* Project:         CMSIS DSP Library    
+* Title:        arm_sin_q31.c    
+*    
+* Description:    Fast sine calculation for Q31 values.   
+*    
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*  
+* Version 1.1.0 2012/02/15 
+*    Updated with more optimizations, bug fixes and minor API changes.  
+*   
+* Version 1.0.10 2011/7/15  
+*    Big Endian support added and Merged M0 and M3/M4 Source code.   
+*    
+* Version 1.0.3 2010/11/29   
+*    Re-organized the CMSIS folders and updated documentation.    
+*     
+* Version 1.0.2 2010/11/11    
+*    Documentation updated.     
+*    
+* Version 1.0.1 2010/10/05     
+*    Production release and review comments incorporated.    
+*    
+* Version 1.0.0 2010/09/20     
+*    Production release and review comments incorporated.    
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**    
+ * @ingroup groupFastMath    
+ */
+
+ /**    
+ * @addtogroup sin    
+ * @{    
+ */
+
+/**   
+ * \par    
+ * Tables generated are in Q31(1.31 Fixed point format)    
+ * Generation of sin values in floating point:    
+ * <pre>tableSize = 256;      
+ * for(n = -1; n < (tableSize + 1); n++)    
+ * {    
+ *    sinTable[n+1]= sin(2*pi*n/tableSize);    
+ * } </pre>    
+ * where pi value is  3.14159265358979    
+ * \par    
+ * Convert Floating point to Q31(Fixed point):    
+ *    (sinTable[i] * pow(2, 31))    
+ * \par    
+ * rounding to nearest integer is done    
+ *     sinTable[i] += (sinTable[i] > 0 ? 0.5 :-0.5);    
+ */
+
+static const q31_t sinTableQ31[259] = {
+  0xfcdbd541, 0x0, 0x3242abf, 0x647d97c, 0x96a9049, 0xc8bd35e, 0xfab272b,
+  0x12c8106f,
+  0x15e21445, 0x18f8b83c, 0x1c0b826a, 0x1f19f97b, 0x2223a4c5, 0x25280c5e,
+  0x2826b928, 0x2b1f34eb,
+  0x2e110a62, 0x30fbc54d, 0x33def287, 0x36ba2014, 0x398cdd32, 0x3c56ba70,
+  0x3f1749b8, 0x41ce1e65,
+  0x447acd50, 0x471cece7, 0x49b41533, 0x4c3fdff4, 0x4ebfe8a5, 0x5133cc94,
+  0x539b2af0, 0x55f5a4d2,
+  0x5842dd54, 0x5a82799a, 0x5cb420e0, 0x5ed77c8a, 0x60ec3830, 0x62f201ac,
+  0x64e88926, 0x66cf8120,
+  0x68a69e81, 0x6a6d98a4, 0x6c242960, 0x6dca0d14, 0x6f5f02b2, 0x70e2cbc6,
+  0x72552c85, 0x73b5ebd1,
+  0x7504d345, 0x7641af3d, 0x776c4edb, 0x78848414, 0x798a23b1, 0x7a7d055b,
+  0x7b5d039e, 0x7c29fbee,
+  0x7ce3ceb2, 0x7d8a5f40, 0x7e1d93ea, 0x7e9d55fc, 0x7f0991c4, 0x7f62368f,
+  0x7fa736b4, 0x7fd8878e,
+  0x7ff62182, 0x7fffffff, 0x7ff62182, 0x7fd8878e, 0x7fa736b4, 0x7f62368f,
+  0x7f0991c4, 0x7e9d55fc,
+  0x7e1d93ea, 0x7d8a5f40, 0x7ce3ceb2, 0x7c29fbee, 0x7b5d039e, 0x7a7d055b,
+  0x798a23b1, 0x78848414,
+  0x776c4edb, 0x7641af3d, 0x7504d345, 0x73b5ebd1, 0x72552c85, 0x70e2cbc6,
+  0x6f5f02b2, 0x6dca0d14,
+  0x6c242960, 0x6a6d98a4, 0x68a69e81, 0x66cf8120, 0x64e88926, 0x62f201ac,
+  0x60ec3830, 0x5ed77c8a,
+  0x5cb420e0, 0x5a82799a, 0x5842dd54, 0x55f5a4d2, 0x539b2af0, 0x5133cc94,
+  0x4ebfe8a5, 0x4c3fdff4,
+  0x49b41533, 0x471cece7, 0x447acd50, 0x41ce1e65, 0x3f1749b8, 0x3c56ba70,
+  0x398cdd32, 0x36ba2014,
+  0x33def287, 0x30fbc54d, 0x2e110a62, 0x2b1f34eb, 0x2826b928, 0x25280c5e,
+  0x2223a4c5, 0x1f19f97b,
+  0x1c0b826a, 0x18f8b83c, 0x15e21445, 0x12c8106f, 0xfab272b, 0xc8bd35e,
+  0x96a9049, 0x647d97c,
+  0x3242abf, 0x0, 0xfcdbd541, 0xf9b82684, 0xf6956fb7, 0xf3742ca2, 0xf054d8d5,
+  0xed37ef91,
+  0xea1debbb, 0xe70747c4, 0xe3f47d96, 0xe0e60685, 0xdddc5b3b, 0xdad7f3a2,
+  0xd7d946d8, 0xd4e0cb15,
+  0xd1eef59e, 0xcf043ab3, 0xcc210d79, 0xc945dfec, 0xc67322ce, 0xc3a94590,
+  0xc0e8b648, 0xbe31e19b,
+  0xbb8532b0, 0xb8e31319, 0xb64beacd, 0xb3c0200c, 0xb140175b, 0xaecc336c,
+  0xac64d510, 0xaa0a5b2e,
+  0xa7bd22ac, 0xa57d8666, 0xa34bdf20, 0xa1288376, 0x9f13c7d0, 0x9d0dfe54,
+  0x9b1776da, 0x99307ee0,
+  0x9759617f, 0x9592675c, 0x93dbd6a0, 0x9235f2ec, 0x90a0fd4e, 0x8f1d343a,
+  0x8daad37b, 0x8c4a142f,
+  0x8afb2cbb, 0x89be50c3, 0x8893b125, 0x877b7bec, 0x8675dc4f, 0x8582faa5,
+  0x84a2fc62, 0x83d60412,
+  0x831c314e, 0x8275a0c0, 0x81e26c16, 0x8162aa04, 0x80f66e3c, 0x809dc971,
+  0x8058c94c, 0x80277872,
+  0x8009de7e, 0x80000000, 0x8009de7e, 0x80277872, 0x8058c94c, 0x809dc971,
+  0x80f66e3c, 0x8162aa04,
+  0x81e26c16, 0x8275a0c0, 0x831c314e, 0x83d60412, 0x84a2fc62, 0x8582faa5,
+  0x8675dc4f, 0x877b7bec,
+  0x8893b125, 0x89be50c3, 0x8afb2cbb, 0x8c4a142f, 0x8daad37b, 0x8f1d343a,
+  0x90a0fd4e, 0x9235f2ec,
+  0x93dbd6a0, 0x9592675c, 0x9759617f, 0x99307ee0, 0x9b1776da, 0x9d0dfe54,
+  0x9f13c7d0, 0xa1288376,
+  0xa34bdf20, 0xa57d8666, 0xa7bd22ac, 0xaa0a5b2e, 0xac64d510, 0xaecc336c,
+  0xb140175b, 0xb3c0200c,
+  0xb64beacd, 0xb8e31319, 0xbb8532b0, 0xbe31e19b, 0xc0e8b648, 0xc3a94590,
+  0xc67322ce, 0xc945dfec,
+  0xcc210d79, 0xcf043ab3, 0xd1eef59e, 0xd4e0cb15, 0xd7d946d8, 0xdad7f3a2,
+  0xdddc5b3b, 0xe0e60685,
+  0xe3f47d96, 0xe70747c4, 0xea1debbb, 0xed37ef91, 0xf054d8d5, 0xf3742ca2,
+  0xf6956fb7, 0xf9b82684,
+  0xfcdbd541, 0x0, 0x3242abf
+};
+
+
+/**   
+ * @brief Fast approximation to the trigonometric sine function for Q31 data.   
+ * @param[in] x Scaled input value in radians.   
+ * @return  sin(x).   
+ *   
+ * The Q31 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*pi), Here range excludes 2*pi.   
+ */
+
+q31_t arm_sin_q31(
+  q31_t x)
+{
+  q31_t sinVal, in, in2;                         /* Temporary variables for input, output */
+  int32_t index;                                 /* Index variables */
+  q31_t wa, wb, wc, wd;                          /* Cubic interpolation coefficients */
+  q31_t a, b, c, d;                              /* Four nearest output values */
+  q31_t *tablePtr;                               /* Pointer to table */
+  q31_t fract, fractCube, fractSquare;           /* Temporary values for fractional values */
+  q31_t oneBy6 = 0x15555555;                     /* Fixed point value of 1/6 */
+  q31_t tableSpacing = TABLE_SPACING_Q31;        /* Table spacing */
+  q31_t temp;                                    /* Temporary variable for intermediate process */
+
+  in = x;
+
+  /* Calculate the nearest index */
+  index = (uint32_t) in / (uint32_t) tableSpacing;
+
+  /* Calculate the nearest value of input */
+  in2 = (q31_t) index *tableSpacing;
+
+  /* Calculation of fractional value */
+  fract = (in - in2) << 8;
+
+  /* fractSquare = fract * fract */
+  fractSquare = ((q31_t) (((q63_t) fract * fract) >> 32));
+  fractSquare = fractSquare << 1;
+
+  /* fractCube = fract * fract * fract */
+  fractCube = ((q31_t) (((q63_t) fractSquare * fract) >> 32));
+  fractCube = fractCube << 1;
+
+  /* Checking min and max index of table */
+  if(index < 0)
+  {
+    index = 0;
+  }
+  else if(index > 256)
+  {
+    index = 256;
+  }
+
+  /* Initialise table pointer */
+  tablePtr = (q31_t *) & sinTableQ31[index];
+
+  /* Cubic interpolation process */
+  /* Calculation of wa */
+  /* wa = -(oneBy6)*fractCube + (fractSquare >> 1u) - (0x2AAAAAAA)*fract; */
+  wa = ((q31_t) (((q63_t) oneBy6 * fractCube) >> 32));
+  temp = 0x2AAAAAAA;
+  wa = (q31_t) ((((q63_t) wa << 32) + ((q63_t) temp * fract)) >> 32);
+  wa = -(wa << 1u);
+  wa += (fractSquare >> 1u);
+
+  /* Read first nearest value of output from the sin table */
+  a = *tablePtr++;
+
+  /* sinVal = a*wa */
+  sinVal = ((q31_t) (((q63_t) a * wa) >> 32));
+
+  /* q31(1.31) Fixed point value of 1 */
+  temp = 0x7FFFFFFF;
+
+  /* Calculation of wb */
+  wb = ((fractCube >> 1u) - (fractSquare + (fract >> 1u))) + temp;
+
+  /* Read second nearest value of output from the sin table */
+  b = *tablePtr++;
+
+  /*  sinVal += b*wb */
+  sinVal = (q31_t) ((((q63_t) sinVal << 32) + (q63_t) b * (wb)) >> 32);
+
+  /* Calculation of wc */
+  wc = -fractCube + fractSquare;
+  wc = (wc >> 1u) + fract;
+
+  /* Read third nearest value of output from the sin table */
+  c = *tablePtr++;
+
+  /*      sinVal += c*wc */
+  sinVal = (q31_t) ((((q63_t) sinVal << 32) + ((q63_t) c * wc)) >> 32);
+
+  /* Calculation of wd */
+  /* wd = (oneBy6) * fractCube - (oneBy6) * fract; */
+  fractCube = fractCube - fract;
+  wd = ((q31_t) (((q63_t) oneBy6 * fractCube) >> 32));
+  wd = (wd << 1u);
+
+  /* Read fourth nearest value of output from the sin table */
+  d = *tablePtr++;
+
+  /* sinVal += d*wd; */
+  sinVal = (q31_t) ((((q63_t) sinVal << 32) + ((q63_t) d * wd)) >> 32);
+
+  /* convert sinVal in 2.30 format to 1.31 format */
+  return (__QADD(sinVal, sinVal));
+
+}
+
+/**    
+ * @} end of sin group    
+ */