About processing the filter and the matrix operation in i.MX6DL.

[keitanagashima]

Dear Sir or madam,

Hello.

My customer's system would like to original process the filter and the matrix operation by CPU power in i.MX6DL.

Is it possible to up the performance by the peripheral of imx6DL(NEON or GPU or IPU)?

If you have these sample, please give me it.

(OS: Linux or other)

Best Regards,

Keita

[CommunityBot · Accepted solution]

This an automatic process.

We are marking this post as solved, due to the either low activity or any reply marked as correct.

If you have additional questions, please create a new post and reference to this closed post.

NXP Community!

[igorpadykov]

Hi Keita

for examples one can look at :

AN4629 Fast Image Processing with i.MX 6 Series

Building OpenCV on ubuntu hardfloat rootfs for i.MX6Q

http://imxcv.blogspot.ru/2014/02/building-opencv-24x-for-freescales-imx6.html

Below description for Cortex-A8 (FFT) example, however I think similar procedure

can be used with Cortex-A9 too.

Fast Fourier Transform (FFT) example (benchmark) benchFFT

can be downloaded from http://www.fftw.org/benchfft/

To compile the benchmark use scratchbox cross compilation tool.

Copy the file to your scratchbox build directory

cp benchfft-3.1.tar.gz /scratchbox/users/jeff/home/jeff/home/.

Extract the package to your build directory in scratchbox

> tar xvzf benchfft-3.1.tar.gz

>cd benchfft-3.1Configure the build for the ARM Cortex-A8 with VFPv3 and NEON co-processors

> ./configure --host=i386 --target=arm --disable-serer CXXFLAGS="-O3 -march=armv7-a -mtune=cortex-a8 -mcpu=cortex-a8 -mfloat-abi=softfp -mfpu=neon -ftree-vectorize -fomit-frame-pointer -ffast-math" CFLAGS="-O3 -march=armv7-a -mtune=cortex-a8 -mcpu=cortex-a8 -mfloat-abi=softfp -mfpu=neon -ftree-vectorize -fomit-frame-pointer -ffast-math" Now compile

>make

Best regards

igor

-----------------------------------------------------------------------------------------------------------------------

Note: If this post answers your question, please click the Correct Answer button. Thank you!

-----------------------------------------------------------------------------------------------------------------------

[keitanagashima]

Hi Igor,

Thank you for your reply.

I got the sample by using GPU.

Do you have sample steps by using IPU or NEON to process the filter and the matrix operation?

BR,

Keita

[igorpadykov]

Hi Keita

I am afraid not.

Best regards

igor