Convolution kernel with ArBB

Convolution kernel with ArBB

Hi there,

Has anyone tried convolution using ArBB. If yes, can I have a code snippet of the ArBB kernel function. It could be a 3x3 averaging kernel operating on 512x512 2D floating iamge data!



4 post / 0 nuovi
Ultimo contenuto
Per informazioni complete sulle ottimizzazioni del compilatore, consultare l'Avviso sull'ottimizzazione

Hello J, ArBB has over 20 different samples, several of being being convolutions. For windows, go to C:\Program Files\Intel\arbb\Beta3\samples\img_processing and you will find some good samples to learn from there. Thanks for your interest.

Thanks Noah for the note.

I'm really interested in benchmarking some image processing algorithms on Intel latest cores using ArBBVs GPUs using CUDA and maybe publish a paper on the findings.

Since, the target of ArBB is to ease off the programmer from low-level programming, these tools have a lot of value for engineers / programmers who like to develop applications with minimal development efforts but giving maximum performance, which indeed saves a lot of thier time and allows them to concentrate more on thier problems, rather than thinking about the parallel implementation.

Please let meknow how can I try ArBB. Please guide me to download page. I'm not interested inthe entire software kit but just want to try out ArBB.


Thanks for being interested in Intel ArBB. It's nice to know you appreciate the value of productivity that ArBB can bring to users. You are welcome to join the beta program. Please go to the page below and click "Sign Up for Beta":

You only need to download an installatio package for the particular platform that you're using, for example, Linux 64-bit. The download size is about 65 MB, which includes the runtime libraries, header files, and sample code.

Intel ArBB works with any standard C++ compiler, such as GCC, MS Visual Studio, or Intel C++ compiler. So you're ready to go if you've already had any of these installed.

Lascia un commento

Eseguire l'accesso per aggiungere un commento. Non siete membri? Iscriviti oggi