ML_MLP_Restart_MPI_4 Test is Failing with Intel 17.4.196 and OpenMPI 2.1.1
Created by: nmhamster
@trilinos/ml - am seeing this error on the latest Intel 17.4.196 build when using OpenMPI 2.1.1. It looks like the precision is set to be very aggressive (small error). Is this a concern for you guys?
526: SA......Total Time = 0.243138
526:
526: ================================================================================
526:
526:
526: *******************************************************
526: ***** Problem: Epetra::CrsMatrix
526: ***** Preconditioned GMRES solution
526: ***** ML (L=3, Jac_pre0/Jac_post0, ~/Amesos_KLU_2)
526: ***** No scaling
526: *******************************************************
526:
526: iter: 0 residual = 1.000000e+00
526: iter: 21 residual = 4.480729e-13
526:
526:
526: Solution time: 0.039518 (sec.)
526: total iterations: 21
526:
526: *******************************************************
526: ***** Problem: Epetra::CrsMatrix
526: ***** Preconditioned GMRES solution
526: ***** ML (L=3, Jac_pre0/Jac_post0, ~/Amesos_KLU_2)
526: ***** No scaling
526: *******************************************************
526:
526: iter: 0 residual = 1.000000e+00
526: iter: 21 residual = 4.480729e-13
526:
526:
526: Solution time: 0.030458 (sec.)
526: total iterations: 21
526: SA......Using 4 processes
526: SA......||x_1 - x_2||_2 = 0
526: SA......Total Time = 0.157298
526:
526: ================================================================================
526:
526:
526: *******************************************************
526: ***** Problem: Epetra::CrsMatrix
526: ***** Preconditioned GMRES solution
526: ***** ML (L=3, Cheby_pre0/Cheby_post0, ~/Amesos_KLU_2)
526: ***** No scaling
526: *******************************************************
526:
526: iter: 0 residual = 1.000000e+00
526: iter: 17 residual = 2.509299e-13
526:
526:
526: Solution time: 0.032598 (sec.)
526: total iterations: 17
526:
526: *******************************************************
526: ***** Problem: Epetra::CrsMatrix
526: ***** Preconditioned GMRES solution
526: ***** ML (L=3, Cheby_pre0/Cheby_post0, ~/Amesos_KLU_2)
526: ***** No scaling
526: *******************************************************
526:
526: iter: 0 residual = 1.000000e+00
526: iter: 17 residual = 2.509299e-13
526:
526:
526: Solution time: 0.027406 (sec.)
526: total iterations: 17
526: SA......Using 4 processes
526: SA......||x_1 - x_2||_2 = 0
526: SA......Total Time = 0.160313
526: Error: `MultiLevelPrecoditioner_Restart.exe' failed!Error: `MultiLevelPrecoditioner_Restart.exe' failed!Error: `MultiLevelPrecoditioner_Restart.exe' failed!
526: ......Total error for exact solution = 2.2261e-14
526:
526: Error: `MultiLevelPrecoditioner_Restart.exe' failed!
526:
526:
526:
526: -------------------------------------------------------
526: Primary job terminated normally, but 1 process returned
526: a non-zero exit code.. Per user-direction, the job has been aborted.
526: -------------------------------------------------------
526: --------------------------------------------------------------------------
526: mpiexec detected that one or more processes exited with non-zero status, thus causing
526: the job to be terminated. The first process to do so was:
526:
526: Process name: [[32722,1],1]
526: Exit code: 1
526: --------------------------------------------------------------------------
1/1 Test #526: ML_MLP_Restart_MPI_4 .............***Failed 2.79 sec