Unexpected behavior while using IPP with sample application

Unexpected behavior while using IPP with sample application

I am new to IPP and am trying to compile one of the sample applications that are freely available to the users, the string processing (pattern matching) application available at:

http://software.intel.com/en-us/articles/code-samples-for-intel-integrat...

My machine has an intel CORE i3 processor - 2 cores and 2 threads for each core. I am using a relatively large file as an input to the pattern matching applicaton for test purposes. What baffles me is that the code runs faster when I have only one logical processor active vs 4 processors (those are the only situations I ran the test under). The way I disable cores is by using msconfig.exe. Could somebody please tell me why this is happening?

The OS is Windows 7 and code is compiled using Intel Parallel Studio 2011.

3 Beiträge / 0 neu
Letzter Beitrag
Nähere Informationen zur Compiler-Optimierung finden Sie in unserem Optimierungshinweis.

Hi Mohammad

I supposed you are checking the ippgrep sample, right?
You may know that not all IPP functions and IPP sample are multi-threaded. As most of ipp string porcessing functions are not threaded and IPP grep sample is not threaded. So there is no much different when run it with 1 or 2 or 4 threading.

You may check a list of the threaded primitives in the IPP library is provided in the ThreadedFunctionsList.txt file located in the library’s doc directory
and the article talking about IPP and threading : http://software.intel.com/en-us/articles/openmp-and-the-intel-ipp-library/

Best Regards,
Ying
http://software.intel.com/en-us/articles/openmp-and-the-intel-ipp-library/

I'd like to add that a list of threaded IPP functions ThreadedFunctionsList.txt is located in a folder:

[ ICCInstallDir ]\Composer XE\Documentation\en_US\Ipp

If you don't have that file please let me know and I'll upload it for you.

Melden Sie sich an, um einen Kommentar zu hinterlassen.