Data race in Pardiso Solver?

[danielsue]

Hi All,

I ran into a tricky problem when test data race for my codes. The Intel Inspector XE 2013 shows that there is data race in the following calling:

282    !C.. Factorization.
283          phase = 22 ! only factorization
284          CALL pardiso (pt, maxfct, mnum, mtype, phase, n, a, ia, ja,     &
285           idum, nrhs, iparm, msglvl, ddum, ddum, error)

But it can still solve the matrix and the result seems correct. How does this data race come from? Would anyone help to check?

Platform: WIN7 X64 Pro + Intel parallel studio XE 2013 (with intel version of Pardiso) + VS2010

I attached the testing codes and matrix data.

Thanks and regards,

Daniel

[Peter_W_Intel]

I'm not the expert of MKL, but cannot repeat this problem.

Tried Fortran Composer 13.0.1.119, and copy mkl example "solverf" sub-directory to my place.

Opened "Intel Composer XE 2013 IA-32 Visual Studio 2010" cmd, cd solverf folder.

Ran "nmake ia2", no problem to build all, cd _results\intel_parallel_is32

inspxe-cl -collect ti3 -- pardiso_unsym_f.exe

0 new problem(s) found.

I used Inspector XE version: Update 6 (build 283349)

[danielsue]

Peter Wang (Intel) wrote:

I'm not the expert of MKL, but cannot repeat this problem.

Tried Fortran Composer 13.0.1.119, and copy mkl example "solverf" sub-directory to my place.

Opened "Intel Composer XE 2013 IA-32 Visual Studio 2010" cmd, cd solverf folder.

Ran "nmake ia2", no problem to build all, cd _results\intel_parallel_is32

inspxe-cl -collect ti3 -- pardiso_unsym_f.exe

0 new problem(s) found.

I used Inspector XE version: Update 6 (build 283349)

Hi Peter,

Please see this forum http://software.intel.com/en-us/forums/topic/392822.

Thanks,

Daniel

[Peter_W_Intel]

Maybe we used different version of sample code, I attached solverf - I used pardiso_unsym_f to test, included Inspector result, no problem detected.

Anyway, you can track your issue with your version of sample code on MKL forum. Thank you.