"Collection failed" on Advanced/General; workaround but annoying

Hi all! I just installed Update 4 of Vtune Amplifier XE 2015 running on Windows 7.

After the update, "Advanced hotspots" and "General Exploration" stopped working (the advanced ones do not work even if I just select "Hotspots"). The exact error message is "Collection failed. The data cannot be displayed" without further info.

Trying to find a better message error, I copied my program inside the "Intel\VTune Amplifier XE 2015\bin64" folder and ran the program directly from the command line using "amplxe-cl -collect advanced-hotspots my_program.exe". It worked!

Difference between C++ code with Blas/Lapack and Matlab

Hi All!

I am trying to implement one Linear programming algorithm in C++. For the matrix multiplication, I use Blas and Lapack. However, I find C++ code performs worse than Matlab as the size of problem is large. Indeed, the difference becomes significant as the size increases. 

I am wondering if it is caused by the optimization tricks of matlab use to call Intel MKL. Could some one help explain why Matlab sometimes outperform C++ with Blas/Lapack?  Is there any way to improve this version of C++ code, or any option to optimize compiling?

Thank you for your time!

ippiResizeSqrPixel_xxx in IPP 9.0 ?

From the compilation errors I get when tyring to port a project to Parallel Studio 2016 (installation package = parallel_studio_xe_2016_setup.exe)

It looks like these (very useful) functions are all gone in IPP 9.0 :


I'll be glad to learn which new/renamed functions one is supposed to use as replacement.

I'm back to Parallel Studio 2015 and IPP 8.0 at the moment.


mkl_ddiamm bug ?

I am trying to multiply two matrices using mkl_ddiamm method:

C = A * B

where A is diagonal matrix 3x3 and B is general matrix 3x3. No matter what I try, i get as a result no A*B, but B*A. This is my sample code. It essentially does SVD decomposition of matrix A and checks, if the computed matrices U, S and VT satisfy all requirements according to theory i.e.

1. U * UT = I, where I us identity matrix

2. V * VT = I

3. U * S * VT = A

Result of temporary operation S * VT is not correct. In fact, the function mkl_ddiamm computes VT * S.

undefined symbol mkl_blas_avx2_cgemm_copyb_ext

I have tried to embed MKL in a java application by embedding the static mkl libraries in a shared lib of mine and then opening that shared lib from the jni. However, I got a link error when I created the shared lib: the symbol mkl_blas_avx2_cgemm_copyb_ext was undefined. I have looked at the static libraries, and indeed, *all* static mkl libraries that have this symbol list them with 'U'. On the other hand, this symbol does not even appear in the shared libraries.

So I'm wondering, why does this symbol appear only in the static libs and why is it not defined?


License not available


I purchased the Parallel Studio XE C++ for OSX on 8/5/2015.  After installing XCode 7.0 I copied the contents of /Library/Application Support/Developer/6.4/XCode/Plug-ins to /Library/Application Support/Developer/7.0/XCode/Plug-ins to have access to the Intel compiler.

I'm now getting a license not available error:

MKL: What to buy?

Hello I am quite disorientated about your MKL library.
I mean, I have developed a calculation software in C# that involves linear system resolution, with dense matrix (complex numbers). My resolution system is very slow, so I was very interested in your library but what to try?
What is Parallel Studio XE? and Which is the difference between Cluster, Professional and Composer Edition?
There are also dll for 64 and 32 bit?

Thank you for any help



Install Crush ; XE 2016 + Xcode 7.0 + OS X 10.10.5

I fail the installation of parallel Studio XE 2016 in my mac pro including OS X* 10.10.5 and Xcode 7.0.

How can I install it successfully?

I tried the click to dmg file  ==> no response 

$ sudo ./install_gui 

=> no response

$ sudo ./ 

./ line 616:  6866 Segmentation fault: 11  "$pset_engine_binary" --TEMP_FOLDER="$temp_folder" $params $root_nonroot 2> /dev/null

Subscribe to Threading