Code::Blocks Forums

User forums => Embedded development => Topic started by: B SRIDEVI on July 10, 2014, 09:36:22 am

Title: performance of codeblocks intel c/c++compiler on intel platform
Post by: B SRIDEVI on July 10, 2014, 09:36:22 am
hi,
  I'm working on Intel atom board where the main constraint in my project  is running time. It has fedora environment.Right now i'm using GNU GCC compiler in codeblocks by linking it with opencv 2.4.8.But i doubt that the performance would be better if i use Intel C/C++ compiler provided by codeblocks IDE since the platform belongs to intel.
can that happen?
If so could u please tell me the settings to be done for using Intel C/C++ compiler in fedora environment,and also the settings required to link it with opencv.
Title: Re: performance of codeblocks intel c/c++compiler on intel platform
Post by: oBFusCATed on July 11, 2014, 12:55:45 am
You need to install the compiler (and probably pay for it) and tell the C::B how to use it.

The second part is goto compiler settings -> select the intel compiler and hit auto detect.
If it doesn't work you'll have to setup paths in the toolchain tab...

Keep in mind that intel-c++ is not that much faster than the newer GCCs, but at the same time it is often full of random bugs.
Also clang is a compiler worth trying.

And to gain full performance benefits you'll probably have to recomile opencv with it.
Title: Re: performance of codeblocks intel c/c++compiler on intel platform
Post by: B SRIDEVI on July 11, 2014, 06:25:47 am
thank you..