Fork of mbed-dsp. CMSIS-DSP library of supporting NEON
Dependents: mbed-os-example-cmsis_dsp_neon
Fork of mbed-dsp by
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。
cmsis_dsp/FastMathFunctions/arm_sin_q15.c
- Committer:
- emilmont
- Date:
- 2012-11-28
- Revision:
- 1:fdd22bb7aa52
- Child:
- 2:da51fb522205
File content as of revision 1:fdd22bb7aa52:
/* ---------------------------------------------------------------------- * Copyright (C) 2010 ARM Limited. All rights reserved. * * $Date: 15. February 2012 * $Revision: V1.1.0 * * Project: CMSIS DSP Library * Title: arm_sin_q15.c * * Description: Fast sine calculation for Q15 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 * Example code for Generation of Q15 Sin Table: * \par * <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 Q15(Fixed point): * (sinTable[i] * pow(2, 15)) * \par * rounding to nearest integer is done * sinTable[i] += (sinTable[i] > 0 ? 0.5 :-0.5); */ static const q15_t sinTableQ15[259] = { 0xfcdc, 0x0, 0x324, 0x648, 0x96b, 0xc8c, 0xfab, 0x12c8, 0x15e2, 0x18f9, 0x1c0c, 0x1f1a, 0x2224, 0x2528, 0x2827, 0x2b1f, 0x2e11, 0x30fc, 0x33df, 0x36ba, 0x398d, 0x3c57, 0x3f17, 0x41ce, 0x447b, 0x471d, 0x49b4, 0x4c40, 0x4ec0, 0x5134, 0x539b, 0x55f6, 0x5843, 0x5a82, 0x5cb4, 0x5ed7, 0x60ec, 0x62f2, 0x64e9, 0x66d0, 0x68a7, 0x6a6e, 0x6c24, 0x6dca, 0x6f5f, 0x70e3, 0x7255, 0x73b6, 0x7505, 0x7642, 0x776c, 0x7885, 0x798a, 0x7a7d, 0x7b5d, 0x7c2a, 0x7ce4, 0x7d8a, 0x7e1e, 0x7e9d, 0x7f0a, 0x7f62, 0x7fa7, 0x7fd9, 0x7ff6, 0x7fff, 0x7ff6, 0x7fd9, 0x7fa7, 0x7f62, 0x7f0a, 0x7e9d, 0x7e1e, 0x7d8a, 0x7ce4, 0x7c2a, 0x7b5d, 0x7a7d, 0x798a, 0x7885, 0x776c, 0x7642, 0x7505, 0x73b6, 0x7255, 0x70e3, 0x6f5f, 0x6dca, 0x6c24, 0x6a6e, 0x68a7, 0x66d0, 0x64e9, 0x62f2, 0x60ec, 0x5ed7, 0x5cb4, 0x5a82, 0x5843, 0x55f6, 0x539b, 0x5134, 0x4ec0, 0x4c40, 0x49b4, 0x471d, 0x447b, 0x41ce, 0x3f17, 0x3c57, 0x398d, 0x36ba, 0x33df, 0x30fc, 0x2e11, 0x2b1f, 0x2827, 0x2528, 0x2224, 0x1f1a, 0x1c0c, 0x18f9, 0x15e2, 0x12c8, 0xfab, 0xc8c, 0x96b, 0x648, 0x324, 0x0, 0xfcdc, 0xf9b8, 0xf695, 0xf374, 0xf055, 0xed38, 0xea1e, 0xe707, 0xe3f4, 0xe0e6, 0xdddc, 0xdad8, 0xd7d9, 0xd4e1, 0xd1ef, 0xcf04, 0xcc21, 0xc946, 0xc673, 0xc3a9, 0xc0e9, 0xbe32, 0xbb85, 0xb8e3, 0xb64c, 0xb3c0, 0xb140, 0xaecc, 0xac65, 0xaa0a, 0xa7bd, 0xa57e, 0xa34c, 0xa129, 0x9f14, 0x9d0e, 0x9b17, 0x9930, 0x9759, 0x9592, 0x93dc, 0x9236, 0x90a1, 0x8f1d, 0x8dab, 0x8c4a, 0x8afb, 0x89be, 0x8894, 0x877b, 0x8676, 0x8583, 0x84a3, 0x83d6, 0x831c, 0x8276, 0x81e2, 0x8163, 0x80f6, 0x809e, 0x8059, 0x8027, 0x800a, 0x8000, 0x800a, 0x8027, 0x8059, 0x809e, 0x80f6, 0x8163, 0x81e2, 0x8276, 0x831c, 0x83d6, 0x84a3, 0x8583, 0x8676, 0x877b, 0x8894, 0x89be, 0x8afb, 0x8c4a, 0x8dab, 0x8f1d, 0x90a1, 0x9236, 0x93dc, 0x9592, 0x9759, 0x9930, 0x9b17, 0x9d0e, 0x9f14, 0xa129, 0xa34c, 0xa57e, 0xa7bd, 0xaa0a, 0xac65, 0xaecc, 0xb140, 0xb3c0, 0xb64c, 0xb8e3, 0xbb85, 0xbe32, 0xc0e9, 0xc3a9, 0xc673, 0xc946, 0xcc21, 0xcf04, 0xd1ef, 0xd4e1, 0xd7d9, 0xdad8, 0xdddc, 0xe0e6, 0xe3f4, 0xe707, 0xea1e, 0xed38, 0xf055, 0xf374, 0xf695, 0xf9b8, 0xfcdc, 0x0, 0x324 }; /** * @brief Fast approximation to the trigonometric sine function for Q15 data. * @param[in] x Scaled input value in radians. * @return sin(x). * * The Q15 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. */ q15_t arm_sin_q15( q15_t x) { q31_t sinVal; /* Temporary variables output */ q15_t *tablePtr; /* Pointer to table */ q15_t fract, in, in2; /* Temporary variables for input, output */ q31_t wa, wb, wc, wd; /* Cubic interpolation coefficients */ q15_t a, b, c, d; /* Four nearest output values */ q15_t fractCube, fractSquare; /* Temporary values for fractional value */ q15_t oneBy6 = 0x1555; /* Fixed point value of 1/6 */ q15_t tableSpacing = TABLE_SPACING_Q15; /* Table spacing */ int32_t index; /* Index variable */ in = x; /* Calculate the nearest index */ index = (int32_t) in / tableSpacing; /* Calculate the nearest value of input */ in2 = (q15_t) ((index) * tableSpacing); /* Calculation of fractional value */ fract = (in - in2) << 8; /* fractSquare = fract * fract */ fractSquare = (q15_t) ((fract * fract) >> 15); /* fractCube = fract * fract * fract */ fractCube = (q15_t) ((fractSquare * fract) >> 15); /* Checking min and max index of table */ if(index < 0) { index = 0; } else if(index > 256) { index = 256; } /* Initialise table pointer */ tablePtr = (q15_t *) & sinTableQ15[index]; /* Cubic interpolation process */ /* Calculation of wa */ /* wa = -(oneBy6)*fractCube + (fractSquare >> 1u) - (0x2AAA)*fract; */ wa = (q31_t) oneBy6 *fractCube; wa += (q31_t) 0x2AAA *fract; wa = -(wa >> 15); wa += ((q31_t) fractSquare >> 1u); /* Read first nearest value of output from the sin table */ a = *tablePtr++; /* sinVal = a * wa */ sinVal = a * wa; /* Calculation of wb */ wb = (((q31_t) fractCube >> 1u) - (q31_t) fractSquare) - (((q31_t) fract >> 1u) - 0x7FFF); /* Read second nearest value of output from the sin table */ b = *tablePtr++; /* sinVal += b*wb */ sinVal += b * wb; /* Calculation of wc */ wc = -(q31_t) fractCube + fractSquare; wc = (wc >> 1u) + fract; /* Read third nearest value of output from the sin table */ c = *tablePtr++; /* sinVal += c*wc */ sinVal += c * wc; /* Calculation of wd */ /* wd = (oneBy6)*fractCube - (oneBy6)*fract; */ fractCube = fractCube - fract; wd = ((q15_t) (((q31_t) oneBy6 * fractCube) >> 15)); /* Read fourth nearest value of output from the sin table */ d = *tablePtr++; /* sinVal += d*wd; */ sinVal += d * wd; /* Convert output value in 1.15(q15) format and saturate */ sinVal = __SSAT((sinVal >> 15), 16); /* Return the output value in 1.15(q15) format */ return ((q15_t) sinVal); } /** * @} end of sin group */