Renesas
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。
Diff: cmsis_dsp/TransformFunctions/arm_rfft_f32.c
- Revision:
- 5:a912b042151f
- Parent:
- 4:9cee975aadce
--- a/cmsis_dsp/TransformFunctions/arm_rfft_f32.c Mon Jun 23 09:30:09 2014 +0100
+++ /dev/null Thu Jan 01 00:00:00 1970 +0000
@@ -1,329 +0,0 @@
-/* ----------------------------------------------------------------------
-* Copyright (C) 2010-2013 ARM Limited. All rights reserved.
-*
-* $Date: 17. January 2013
-* $Revision: V1.4.1
-*
-* Project: CMSIS DSP Library
-* Title: arm_rfft_f32.c
-*
-* Description: RFFT & RIFFT Floating point process function
-*
-* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
-*
-* Redistribution and use in source and binary forms, with or without
-* modification, are permitted provided that the following conditions
-* are met:
-* - Redistributions of source code must retain the above copyright
-* notice, this list of conditions and the following disclaimer.
-* - Redistributions in binary form must reproduce the above copyright
-* notice, this list of conditions and the following disclaimer in
-* the documentation and/or other materials provided with the
-* distribution.
-* - Neither the name of ARM LIMITED nor the names of its contributors
-* may be used to endorse or promote products derived from this
-* software without specific prior written permission.
-*
-* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
-* FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
-* COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
-* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
-* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
-* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
-* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
-* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
-* ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
-* POSSIBILITY OF SUCH DAMAGE.
-* -------------------------------------------------------------------- */
-
-#include "arm_math.h"
-
-extern void arm_radix4_butterfly_f32(
- float32_t * pSrc,
- uint16_t fftLen,
- float32_t * pCoef,
- uint16_t twidCoefModifier);
-
-extern void arm_radix4_butterfly_inverse_f32(
- float32_t * pSrc,
- uint16_t fftLen,
- float32_t * pCoef,
- uint16_t twidCoefModifier,
- float32_t onebyfftLen);
-
-extern void arm_bitreversal_f32(
- float32_t * pSrc,
- uint16_t fftSize,
- uint16_t bitRevFactor,
- uint16_t * pBitRevTab);
-
-/**
- * @ingroup groupTransforms
- */
-
-/*--------------------------------------------------------------------
- * Internal functions prototypes
- *--------------------------------------------------------------------*/
-
-void arm_split_rfft_f32(
- float32_t * pSrc,
- uint32_t fftLen,
- float32_t * pATable,
- float32_t * pBTable,
- float32_t * pDst,
- uint32_t modifier);
-void arm_split_rifft_f32(
- float32_t * pSrc,
- uint32_t fftLen,
- float32_t * pATable,
- float32_t * pBTable,
- float32_t * pDst,
- uint32_t modifier);
-
-/**
- * @addtogroup RealFFT
- * @{
- */
-
-/**
- * @brief Processing function for the floating-point RFFT/RIFFT.
- * @deprecated Do not use this function. It has been superceded by \ref arm_rfft_fast_f32 and will be removed
- * in the future.
- * @param[in] *S points to an instance of the floating-point RFFT/RIFFT structure.
- * @param[in] *pSrc points to the input buffer.
- * @param[out] *pDst points to the output buffer.
- * @return none.
- */
-
-void arm_rfft_f32(
- const arm_rfft_instance_f32 * S,
- float32_t * pSrc,
- float32_t * pDst)
-{
- const arm_cfft_radix4_instance_f32 *S_CFFT = S->pCfft;
-
-
- /* Calculation of Real IFFT of input */
- if(S->ifftFlagR == 1u)
- {
- /* Real IFFT core process */
- arm_split_rifft_f32(pSrc, S->fftLenBy2, S->pTwiddleAReal,
- S->pTwiddleBReal, pDst, S->twidCoefRModifier);
-
-
- /* Complex radix-4 IFFT process */
- arm_radix4_butterfly_inverse_f32(pDst, S_CFFT->fftLen,
- S_CFFT->pTwiddle,
- S_CFFT->twidCoefModifier,
- S_CFFT->onebyfftLen);
-
- /* Bit reversal process */
- if(S->bitReverseFlagR == 1u)
- {
- arm_bitreversal_f32(pDst, S_CFFT->fftLen,
- S_CFFT->bitRevFactor, S_CFFT->pBitRevTable);
- }
- }
- else
- {
-
- /* Calculation of RFFT of input */
-
- /* Complex radix-4 FFT process */
- arm_radix4_butterfly_f32(pSrc, S_CFFT->fftLen,
- S_CFFT->pTwiddle, S_CFFT->twidCoefModifier);
-
- /* Bit reversal process */
- if(S->bitReverseFlagR == 1u)
- {
- arm_bitreversal_f32(pSrc, S_CFFT->fftLen,
- S_CFFT->bitRevFactor, S_CFFT->pBitRevTable);
- }
-
-
- /* Real FFT core process */
- arm_split_rfft_f32(pSrc, S->fftLenBy2, S->pTwiddleAReal,
- S->pTwiddleBReal, pDst, S->twidCoefRModifier);
- }
-
-}
-
-/**
- * @} end of RealFFT group
- */
-
-/**
- * @brief Core Real FFT process
- * @param[in] *pSrc points to the input buffer.
- * @param[in] fftLen length of FFT.
- * @param[in] *pATable points to the twiddle Coef A buffer.
- * @param[in] *pBTable points to the twiddle Coef B buffer.
- * @param[out] *pDst points to the output buffer.
- * @param[in] modifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table.
- * @return none.
- */
-
-void arm_split_rfft_f32(
- float32_t * pSrc,
- uint32_t fftLen,
- float32_t * pATable,
- float32_t * pBTable,
- float32_t * pDst,
- uint32_t modifier)
-{
- uint32_t i; /* Loop Counter */
- float32_t outR, outI; /* Temporary variables for output */
- float32_t *pCoefA, *pCoefB; /* Temporary pointers for twiddle factors */
- float32_t CoefA1, CoefA2, CoefB1; /* Temporary variables for twiddle coefficients */
- float32_t *pDst1 = &pDst[2], *pDst2 = &pDst[(4u * fftLen) - 1u]; /* temp pointers for output buffer */
- float32_t *pSrc1 = &pSrc[2], *pSrc2 = &pSrc[(2u * fftLen) - 1u]; /* temp pointers for input buffer */
-
- /* Init coefficient pointers */
- pCoefA = &pATable[modifier * 2u];
- pCoefB = &pBTable[modifier * 2u];
-
- i = fftLen - 1u;
-
- while(i > 0u)
- {
- /*
- outR = (pSrc[2 * i] * pATable[2 * i] - pSrc[2 * i + 1] * pATable[2 * i + 1]
- + pSrc[2 * n - 2 * i] * pBTable[2 * i] +
- pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]);
- */
-
- /* outI = (pIn[2 * i + 1] * pATable[2 * i] + pIn[2 * i] * pATable[2 * i + 1] +
- pIn[2 * n - 2 * i] * pBTable[2 * i + 1] -
- pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); */
-
- /* read pATable[2 * i] */
- CoefA1 = *pCoefA++;
- /* pATable[2 * i + 1] */
- CoefA2 = *pCoefA;
-
- /* pSrc[2 * i] * pATable[2 * i] */
- outR = *pSrc1 * CoefA1;
- /* pSrc[2 * i] * CoefA2 */
- outI = *pSrc1++ * CoefA2;
-
- /* (pSrc[2 * i + 1] + pSrc[2 * fftLen - 2 * i + 1]) * CoefA2 */
- outR -= (*pSrc1 + *pSrc2) * CoefA2;
- /* pSrc[2 * i + 1] * CoefA1 */
- outI += *pSrc1++ * CoefA1;
-
- CoefB1 = *pCoefB;
-
- /* pSrc[2 * fftLen - 2 * i + 1] * CoefB1 */
- outI -= *pSrc2-- * CoefB1;
- /* pSrc[2 * fftLen - 2 * i] * CoefA2 */
- outI -= *pSrc2 * CoefA2;
-
- /* pSrc[2 * fftLen - 2 * i] * CoefB1 */
- outR += *pSrc2-- * CoefB1;
-
- /* write output */
- *pDst1++ = outR;
- *pDst1++ = outI;
-
- /* write complex conjugate output */
- *pDst2-- = -outI;
- *pDst2-- = outR;
-
- /* update coefficient pointer */
- pCoefB = pCoefB + (modifier * 2u);
- pCoefA = pCoefA + ((modifier * 2u) - 1u);
-
- i--;
-
- }
-
- pDst[2u * fftLen] = pSrc[0] - pSrc[1];
- pDst[(2u * fftLen) + 1u] = 0.0f;
-
- pDst[0] = pSrc[0] + pSrc[1];
- pDst[1] = 0.0f;
-
-}
-
-
-/**
- * @brief Core Real IFFT process
- * @param[in] *pSrc points to the input buffer.
- * @param[in] fftLen length of FFT.
- * @param[in] *pATable points to the twiddle Coef A buffer.
- * @param[in] *pBTable points to the twiddle Coef B buffer.
- * @param[out] *pDst points to the output buffer.
- * @param[in] modifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table.
- * @return none.
- */
-
-void arm_split_rifft_f32(
- float32_t * pSrc,
- uint32_t fftLen,
- float32_t * pATable,
- float32_t * pBTable,
- float32_t * pDst,
- uint32_t modifier)
-{
- float32_t outR, outI; /* Temporary variables for output */
- float32_t *pCoefA, *pCoefB; /* Temporary pointers for twiddle factors */
- float32_t CoefA1, CoefA2, CoefB1; /* Temporary variables for twiddle coefficients */
- float32_t *pSrc1 = &pSrc[0], *pSrc2 = &pSrc[(2u * fftLen) + 1u];
-
- pCoefA = &pATable[0];
- pCoefB = &pBTable[0];
-
- while(fftLen > 0u)
- {
- /*
- outR = (pIn[2 * i] * pATable[2 * i] + pIn[2 * i + 1] * pATable[2 * i + 1] +
- pIn[2 * n - 2 * i] * pBTable[2 * i] -
- pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1]);
-
- outI = (pIn[2 * i + 1] * pATable[2 * i] - pIn[2 * i] * pATable[2 * i + 1] -
- pIn[2 * n - 2 * i] * pBTable[2 * i + 1] -
- pIn[2 * n - 2 * i + 1] * pBTable[2 * i]);
-
- */
-
- CoefA1 = *pCoefA++;
- CoefA2 = *pCoefA;
-
- /* outR = (pSrc[2 * i] * CoefA1 */
- outR = *pSrc1 * CoefA1;
-
- /* - pSrc[2 * i] * CoefA2 */
- outI = -(*pSrc1++) * CoefA2;
-
- /* (pSrc[2 * i + 1] + pSrc[2 * fftLen - 2 * i + 1]) * CoefA2 */
- outR += (*pSrc1 + *pSrc2) * CoefA2;
-
- /* pSrc[2 * i + 1] * CoefA1 */
- outI += (*pSrc1++) * CoefA1;
-
- CoefB1 = *pCoefB;
-
- /* - pSrc[2 * fftLen - 2 * i + 1] * CoefB1 */
- outI -= *pSrc2-- * CoefB1;
-
- /* pSrc[2 * fftLen - 2 * i] * CoefB1 */
- outR += *pSrc2 * CoefB1;
-
- /* pSrc[2 * fftLen - 2 * i] * CoefA2 */
- outI += *pSrc2-- * CoefA2;
-
- /* write output */
- *pDst++ = outR;
- *pDst++ = outI;
-
- /* update coefficient pointer */
- pCoefB = pCoefB + (modifier * 2u);
- pCoefA = pCoefA + ((modifier * 2u) - 1u);
-
- /* Decrement loop count */
- fftLen--;
- }
-
-}