I use gfortran 4.4.1 (and ifort 11.0) with acml 4.3.0 in Ubuntu 9.10 32bit (and Scientific Linux 5.3 32bit). I compile this code:

      program eigenlapack      

      PARAMETER        (n=1000)

      PARAMETER        (lwork=8*n)      

      integer           m, info, iwork(5*n), IL, IU, ifail(n),i,j

      REAL              w(n), work(lwork), a(n,n), VL, VU, z(n,n)

      open(11,file='eigen.txt',form='formatted',status='UNKNOWN')

      IL = 1

      IU = 5

      do 10 i=1,n

        do 20 j=1,n

          a(i,j)=rand()*100

   20 continue

   10 continue

!$omp parallel default(firstprivate)

      CALL SSYEVX ('N', 'A',  'U', n, a, n, VL, VU, IL, IU, 0.0d0, 

     1              m, w, z, n, work, lwork,  iwork, ifail, info)

      write(*,'(3x,a)') 'Number of eigenvalues:'

      write(*,'(3x,I100)') m

      write(*,'(3x,a)') 'Eigenvalues:'

      write(11,'(10000f12.2)') (w(i),i=1,n)

!$omp end parallel

      end

If I set "export OMP_NUM_THREADS=1" then it works correctly. But if I use more than one thread (OMP_NUM_THREADS=2) then it don't work:

Segmentation fault (core dumped)

It work with variable n=416 or less. 

Single version of code work right.

What can I do?

The bug is easily reproduced.  It happens with a 64-bit OS also, and when linking either with the single threaded or OpenMP ACML library. I used ACML 4.3.0 gfortran64 and gcc/gfortran 4.4.1 to reproduce the problem.

GDB quickly reveals that the segfault occurs on the openmp parallel clause, and looks immediately like a stack size issue.  A quick search in google turned up similar issues.

I resolved the problem like this:  Change the declaration for the REAL variables to:

REAL,save::     w(n), work(lwork), a(n,n), VL, VU, z(n,n)

I think this is moving these larger arrays off of the stack and into allocated memory.  You could also solve the problem by increasing the available stack size.

Such a simple change, but it fixed the problem for me.