Problem with the latest version of Cluster Studio XE 2013 for Windows

[www_q_]

I just bought Intel Cluster Studio 2013 XE for Windows, I used the SN to login to Intel's website to download the latest package (Cluster Studio XE 2013 SP1 Update 1 for windows, built in Feb. 2014).

Following the installation instructions, I installed all the packages here (including single and multi-threaded static libaries for IPP).

The problem I met is (with Microsoft visual studio):

The integration with MS visual studio 2012 is bugged: IPP simply didnt work:

(1) It give me unsolved symbol errors even if I enable IPP in VS2012 (I have used Parallel Studio 2013 before, so dont assume I dont know how to configure IPP in VS2012), therefore I have to manually add the necessary ipp lib files to the linker to build the binary;

(2), It seems that IPP will always use DLL libaries even if I set it to use ipp static libs in VS2012.

A evaluation version of Paralell Studio 2013 XE Update 4 and the IPP within it (I think it is version 7.11?) works perfectly with VS2012 and I have not tested any other functionalities of cluster studios, I dont know, if it is  a problem of (1) IPP 8.1 or (2)CS2013 SP1 Update 1 (3)Both.

Btw, if there any tech support hotline in China for Intel software?

[Pavel_B_Intel1]

Hi,

I have the same configuration:

• MSVC2012
• Intel Parallel Studio 2013

I build simple "Hello world" example without linking IPP libraries and build an example which uses IPP libraries without problem.

To configure how to link IPP libraries you should look at the project properties -> Configuration Properties -> Intel Performance Libraries. In this tab you can manage how to link or don't link IPP libraries by using options in "Use Intel IPP" field. Please set "no" if you  don't wont to use IPP libraries or set "Single-threaded Static Library" if you would like to use IPP libraries.

Why I propose to use single-threaded libraries:

• IPP dynamic libraries are huge size
• multi-threaded libraries are deprecated and more effective to parallel pipeline from several IPP functions + user code inside user application instead of parallel small part of code inside an IPP function.

Pavel