[Trilinos-Users] Seg fault in AZ_manage_memory

[Heroux, Michael A]

Glen,

Might you be going past the address range that int supports (2^31)?

If so, this is not a graceful way for AztecOO to tell you, so let me know.

Mike


On 9/30/10 4:19 PM, "Glen Hansen" <glen.hansen at inl.gov> wrote:

> All,
> 
> I am seeing a seg fault on a pretty large problem that is showing up in
> AZ_manage_memory.
> 
> I am using ML as an approximate preconditioner inversion method and passing
> the interface to my preconditioner to NOX::Epetra::LinearSystemAztecOO. I get
> the following traceback from gdb:
> 
> Program received signal EXC_BAD_ACCESS, Could not access memory.
> Reason: KERN_INVALID_ADDRESS at address: 0x00000000d3208ca8
> 0x00000001003b69d8 in AZ_manage_memory (input_size=2590485664, action=0,
> type=-914901, name=0x7fff5fbfd2d0 "vblock in gmres0", status=0x7fff5fbfd36c)
> at /Users/hansga/Codes/Trilinos/packages/aztecoo/src/az_util.c:991
> 991       temp->name = (char *)     &(dtmp[(aligned_str_mem+aligned_size)/
> (gdb) where
> #0  0x00000001003b69d8 in AZ_manage_memory (input_size=2590485664, action=0,
> type=-914901, name=0x7fff5fbfd2d0 "vblock in gmres0", status=0x7fff5fbfd36c)
> at /Users/hansga/Codes/Trilinos/packages/aztecoo/src/az_util.c:991
> #1  0x00000001003c89f4 in AZ_pgmres (b=0x1138d2000, x=0x114526000,
> weight=0x101124ed8, options=0x101124dd0, params=0x101124e90,
> proc_config=0x101124d60, status=0x101124f80, Amat=0x10112cf70,
> precond=0x10112d9e0, convergence_info=0x10112dd10) at
> /Users/hansga/Codes/Trilinos/packages/aztecoo/src/az_gmres.c:232
> #2  0x00000001003e09af in AZ_oldsolve (x=0x114526000, b=0x1138d2000,
> options=0x101124dd0, params=0x101124e90, status=0x101124f80,
> proc_config=0x101124d60, Amat=0x10112cf70, precond=0x10112d9e0,
> scaling=0x10112dce0) at
> /Users/hansga/Codes/Trilinos/packages/aztecoo/src/az_solve.c:540
> #3  0x00000001003dfd6b in AZ_iterate (x=0x114526000, b=0x1138d2000,
> options=0x101124dd0, params=0x101124e90, status=0x101124f80,
> proc_config=0x101124d60, Amat=0x10112cf70, precond=0x10112d9e0, scaling=0x0)
> at /Users/hansga/Codes/Trilinos/packages/aztecoo/src/az_solve.c:190
> #4  0x000000010039b212 in AztecOO::Iterate (this=0x101124ab0, MaxIters=400,
> Tolerance=0.0001) at
> /Users/hansga/Codes/Trilinos/packages/aztecoo/src/AztecOO.cpp:912
> #5  0x00000001000a9361 in
> NOX::Epetra::LinearSystemAztecOO::applyJacobianInverse (this=0x1011246d0,
> p=@0x108a9ad38, input=@0x101127510, result=@0x101127ea0) at
> /Users/hansga/Codes/Trilinos/packages/nox/src-epetra/NOX_Epetra_LinearSystem_A
> ztecOO.C:686
> #6  0x000000010009d934 in NOX::Epetra::Group::applyJacobianInverse
> (this=0x101126f30, p=@0x108a9ad38, input=@0x101127510, result=@0x101127ea0) at
> /Users/hansga/Codes/Trilinos/packages/nox/src-epetra/NOX_Epetra_Group.C:429
> #7  0x000000010009d2e0 in NOX::Epetra::Group::computeNewton (this=0x101126f30,
> p=@0x108a9ad38) at
> /Users/hansga/Codes/Trilinos/packages/nox/src-epetra/NOX_Epetra_Group.C:353
> #8  0x00000001000e0d19 in NOX::Direction::Newton::compute (this=0x10112c9f0,
> dir=@0x10112c430, soln=@0x101126f30, solver=@0x10112aa70) at
> /Users/hansga/Codes/Trilinos/packages/nox/src/NOX_Direction_Newton.C:127
> #9  0x00000001000d83b2 in NOX::Direction::Generic::compute (this=0x10112c9f0,
> d=@0x10112c430, g=@0x101126f30, s=@0x10112aa70) at
> /Users/hansga/Codes/Trilinos/packages/nox/src/NOX_Direction_Generic.C:50
> #10 0x00000001000e10c4 in NOX::Direction::Newton::compute (this=0x10112c9f0,
> dir=@0x10112c430, soln=@0x101126f30, solver=@0x10112aa70) at
> /Users/hansga/Codes/Trilinos/packages/nox/src/NOX_Direction_Newton.C:153
> #11 0x00000001000f1dca in NOX::Solver::LineSearchBased::step
> (this=0x10112aa70) at
> /Users/hansga/Codes/Trilinos/packages/nox/src/NOX_Solver_LineSearchBased.C:165
> #12 0x00000001000f2198 in NOX::Solver::LineSearchBased::solve
> (this=0x10112aa70) at
> /Users/hansga/Codes/Trilinos/packages/nox/src/NOX_Solver_LineSearchBased.C:224
> #13 0x000000010002c7a8 in main (argc=2, argv=0x7fff5fbfeee8) at Mesh.C:335
> Current language:  auto; currently c
> (gdb) 
> 
> Any thoughts on what might be happening to me?
> 
> Thanks!
> 
> Glen
> 
> 
> -
> 
> **************************************************
> Glen Hansen                               (208) 526-8494
> Idaho National Laboratory
> MS 3840                                                EROB 320
> **************************************************
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Trilinos-Users mailing list
> Trilinos-Users at software.sandia.gov
> http://software.sandia.gov/mailman/listinfo/trilinos-users