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_sqrt_q15.c
- Committer:
- emilmont
- Date:
- 2012-11-28
- Revision:
- 1:fdd22bb7aa52
- Child:
- 2:da51fb522205
File content as of revision 1:fdd22bb7aa52:
/* ---------------------------------------------------------------------- * Copyright (C) 2011 ARM Limited. All rights reserved. * * $Date: 15. February 2012 * $Revision: V1.1.0 * * Project: CMSIS DSP Library * Title: arm_sqrt_q15.c * * Description: Q15 square root function. * * 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 1.0.0 2011/03/08 * Alpha release. * * Version 1.0.1 2011/09/30 * Beta release. * * -------------------------------------------------------------------- */ #include "arm_math.h" #include "arm_common_tables.h" /** * @ingroup groupFastMath */ /** * @addtogroup SQRT * @{ */ /** * @brief Q15 square root function. * @param[in] in input value. The range of the input value is [0 +1) or 0x0000 to 0x7FFF. * @param[out] *pOut square root of input value. * @return The function returns ARM_MATH_SUCCESS if input value is positive value or ARM_MATH_ARGUMENT_ERROR if * <code>in</code> is negative value and returns zero output for negative values. */ arm_status arm_sqrt_q15( q15_t in, q15_t * pOut) { q15_t number, temp1, var1, signBits1, half; q31_t bits_val1; float32_t temp_float1; number = in; /* If the input is a positive number then compute the signBits. */ if(number > 0) { signBits1 = __CLZ(number) - 17; /* Shift by the number of signBits1 */ if((signBits1 % 2) == 0) { number = number << signBits1; } else { number = number << (signBits1 - 1); } /* Calculate half value of the number */ half = number >> 1; /* Store the number for later use */ temp1 = number; /*Convert to float */ temp_float1 = number * 3.051757812500000e-005f; /*Store as integer */ bits_val1 = *(int *) &temp_float1; /* Subtract the shifted value from the magic number to give intial guess */ bits_val1 = 0x5f3759df - (bits_val1 >> 1); // gives initial guess /* Store as float */ temp_float1 = *(float *) &bits_val1; /* Convert to integer format */ var1 = (q31_t) (temp_float1 * 16384); /* 1st iteration */ var1 = ((q15_t) ((q31_t) var1 * (0x3000 - ((q15_t) ((((q15_t) (((q31_t) var1 * var1) >> 15)) * (q31_t) half) >> 15))) >> 15)) << 2; /* 2nd iteration */ var1 = ((q15_t) ((q31_t) var1 * (0x3000 - ((q15_t) ((((q15_t) (((q31_t) var1 * var1) >> 15)) * (q31_t) half) >> 15))) >> 15)) << 2; /* 3rd iteration */ var1 = ((q15_t) ((q31_t) var1 * (0x3000 - ((q15_t) ((((q15_t) (((q31_t) var1 * var1) >> 15)) * (q31_t) half) >> 15))) >> 15)) << 2; /* Multiply the inverse square root with the original value */ var1 = ((q15_t) (((q31_t) temp1 * var1) >> 15)) << 1; /* Shift the output down accordingly */ if((signBits1 % 2) == 0) { var1 = var1 >> (signBits1 / 2); } else { var1 = var1 >> ((signBits1 - 1) / 2); } *pOut = var1; return (ARM_MATH_SUCCESS); } /* If the number is a negative number then store zero as its square root value */ else { *pOut = 0; return (ARM_MATH_ARGUMENT_ERROR); } } /** * @} end of SQRT group */