From: Andrea Hawkins <andjhawkins@gm...>  20100308 20:43:22

Hello I'm attempting to do mesh refinement with the PatchRecoveryErrorEstimator (both coarsening and refining). This works quite well, and if I do a restart with the adapted mesh just asking it to continue with refining it works fine. However, if I ask it to continue both refining and coarsening I get the following error when it attempts to refine: [11]PETSC ERROR: [10]PETSC ERROR: [8]PETSC ERROR:  [11]PETSC ERROR: Caught signal number 7 BUS: Bus Error, possibly illegal memory access   [11]PETSC ERROR: [10]PETSC ERROR: Try option start_in_debugger or on_error_attach_debugger [8]PETSC ERROR: [11]PETSC ERROR: or see http://www.mcs.anl.gov/petsc/petscas/documentation/troubleshooting.html#SignalCaught signal number 7 BUS: Bus Error, possibly illegal memory access Caught signal number 7 BUS: Bus Error, possibly illegal memory access [11]PETSC ERROR: [8]PETSC ERROR: [10]PETSC ERROR: or try http://valgrind.org on linux or man libgmalloc on Apple to find memory corruption errors Try option start_in_debugger or on_error_attach_debugger Try option start_in_debugger or on_error_attach_debugger [11]PETSC ERROR: [8]PETSC ERROR: [10]PETSC ERROR: configure using withdebugging=yes, recompile, link, and run or see http://www.mcs.anl.gov/petsc/petscas/documentation/troubleshooting.html#Signalor see http://www.mcs.anl.gov/petsc/petscas/documentation/troubleshooting.html#Signal[11]PETSC ERROR: [8]PETSC ERROR: [10]PETSC ERROR: to get more information on the crash. or try http://valgrind.org on linux or man libgmalloc on Apple to find memory corruption errors or try http://valgrind.org on linux or man libgmalloc on Apple to find memory corruption errors [10]PETSC ERROR: [8]PETSC ERROR: configure using withdebugging=yes, recompile, link, and run configure using withdebugging=yes, recompile, link, and run [11]PETSC ERROR: [8]PETSC ERROR: [10]PETSC ERROR: to get more information on the crash. to get more information on the crash.  Error Message  [8]PETSC ERROR: [11]PETSC ERROR:  Error Message  [10]PETSC ERROR: Signal received! [8]PETSC ERROR:  Error Message  [10]PETSC ERROR: Signal received! Signal received! [8]PETSC ERROR: [11]PETSC ERROR: [10]PETSC ERROR:   [8]PETSC ERROR:  [11]PETSC ERROR: [10]PETSC ERROR: Petsc Release Version 2.3.3, Patch 6, Fri Sep 21 18:02:04 CDT 2007 HG revision: cd8f2b29cffd6721068fcaa6a9b635012be15ca2 Petsc Release Version 2.3.3, Patch 6, Fri Sep 21 18:02:04 CDT 2007 HG revision: cd8f2b29cffd6721068fcaa6a9b635012be15ca2 [11]PETSC ERROR: [10]PETSC ERROR: See docs/changes/index.html for recent updates. See docs/changes/index.html for recent updates. Petsc Release Version 2.3.3, Patch 6, Fri Sep 21 18:02:04 CDT 2007 HG revision: cd8f2b29cffd6721068fcaa6a9b635012be15ca2 [11]PETSC ERROR: [10]PETSC ERROR: See docs/faq.html for hints about trouble shooting. See docs/faq.html for hints about trouble shooting. [11]PETSC ERROR: [10]PETSC ERROR: See docs/index.html for manual pages. [8]PETSC ERROR: [11]PETSC ERROR: See docs/changes/index.html for recent updates.  See docs/index.html for manual pages. [8]PETSC ERROR: [11]PETSC ERROR: [10]PETSC ERROR: See docs/faq.html for hints about trouble shooting. /home/utexas/ti/ahawkins/GenAlpha/gen_alphaopt on a em64tcxx named c11305 by ahawkins Mon Mar 8 13:15:39 2010  [8]PETSC ERROR: [11]PETSC ERROR: [10]PETSC ERROR: See docs/index.html for manual pages. Libraries linked from /opt/apps/intel10/mvapich1/petsc/2.3.3/lib/em64tcxx /home/utexas/ti/ahawkins/GenAlpha/gen_alphaopt on a em64tcxx named c11305 by ahawkins Mon Mar 8 13:15:39 2010 .............. Has anyone ever seen this before? Thanks, Andrea 
From: Andrea Hawkins <andjhawkins@gm...>  20100308 21:22:50

Ok... So.... Through some combination of make clean and make the problem appears to be fixed... And I'm having trouble reproducing it. Strange... If I can figure out what I did to make that happen again, I will be sure to get the traceback. Sorry about that. Andrea On Mon, Mar 8, 2010 at 2:56 PM, Roy Stogner <roystgnr@...> wrote: > > On Mon, 8 Mar 2010, Andrea Hawkins wrote: > >> I'm attempting to do mesh refinement with the >> PatchRecoveryErrorEstimator (both coarsening and refining). This works >> quite well, and if I do a restart with the adapted mesh just asking it >> to continue with refining it works fine. However, if I ask it to >> continue both refining and coarsening I get the following error when >> it attempts to refine: > >> [11]PETSC ERROR: Caught signal number 7 BUS: Bus Error, possibly >> illegal memory access > >> Has anyone ever seen this before? > > Vikram ran into some asyetunidentified problem with the new L2 > option for PatchRecovery last week, but that's almost certainly > unrelated. > > There's a bug in our libHilbert interface that could have corrupted an > old output file; that's why for now we disable libHilbert by default. > Assuming you're trying to restart from newly generated output files, > would you make sure LIBMESH_HAVE_LIBHILBERT is commented out in > libmesh_config.h, please? > > Is this the first error you hit in dbg mode? If not then knowing what > assert first fails might help. If so, then could you run though > start_in_debugger and give us a stack trace? > > Thanks, >  > Roy > 
From: Roy Stogner <roystgnr@ic...>  20100308 21:26:45

On Mon, 8 Mar 2010, Andrea Hawkins wrote: > I'm attempting to do mesh refinement with the > PatchRecoveryErrorEstimator (both coarsening and refining). This works > quite well, and if I do a restart with the adapted mesh just asking it > to continue with refining it works fine. However, if I ask it to > continue both refining and coarsening I get the following error when > it attempts to refine: > [11]PETSC ERROR: Caught signal number 7 BUS: Bus Error, possibly > illegal memory access > Has anyone ever seen this before? Vikram ran into some asyetunidentified problem with the new L2 option for PatchRecovery last week, but that's almost certainly unrelated. There's a bug in our libHilbert interface that could have corrupted an old output file; that's why for now we disable libHilbert by default. Assuming you're trying to restart from newly generated output files, would you make sure LIBMESH_HAVE_LIBHILBERT is commented out in libmesh_config.h, please? Is this the first error you hit in dbg mode? If not then knowing what assert first fails might help. If so, then could you run though start_in_debugger and give us a stack trace? Thanks,  Roy 
From: Roy Stogner <roystgnr@ic...>  20100308 21:41:53

On Mon, 8 Mar 2010, Andrea Hawkins wrote: > Ok... So.... Through some combination of make clean and make the > problem appears to be fixed... And I'm having trouble reproducing it. > > Strange... If I can figure out what I did to make that happen again, I > will be sure to get the traceback. Sorry about that. It's happened to me too. One possible problem: if you're not using a .dependlike system (or you're using an outofdate .depend file) in your application build directory, and if you svn update the library underneath it, make may not rebuild every object file that needs it. Then you have mismatching struct definitions or function signatures, and either a member variable or a stack variable is misaligned. We've talked about no longer caching dependency information, and getting rid of this problem for good, but: 1. That can slow down make on highlatency filesystems 2. We're lazy  Roy 
From: Liang <goeasyon@gm...>  20100321 03:17:00

Hi developers and users, Seems I sent to wrong address last email, sorry for that. I tried to expand the tensor form of NS to the general PDEs as shown in the attached picture, is this same to the governing equation that ex13 used? please correct me if i am wrong. Is any book introduce deriving the nonlinear PDEs to the element equations? nonlinear terms in the equation bother me a lot, I don't know how to derive their weak forms. Another question is about the constant \theta in the ex13, which is equal to one in the equation, the description goes like: "Find out what the timestep size parameter is from the system, and the value of theta for the theta method. We use implicit Euler (theta=1) for this simulation even though it is only firstorder accurate in time. " can I understand it as the coordinator for the CrankNicolson method, theta =0.5 means the classic CrankNicolson, theta = 0/1 means pure forward/backward Euler method? There are no \mu and \rho in the assemble equation, so both parameters are evaluated one? Many thanks ! Liang ===============copy from ex13 ================== for (unsigned int i=0; i<n_u_dofs; i++) { Fu(i) += JxW[qp]*(u_old*phi[i][qp]  // massmatrix term (1.theta)*dt*(U_old*grad_u_old)*phi[i][qp] + // convection term (1.theta)*dt*p_old*dphi[i][qp](0)  // pressure term on rhs (1.theta)*dt*(grad_u_old*dphi[i][qp]) + // diffusion term on rhs theta*dt*(U*grad_u)*phi[i][qp]); // Newton term Fv(i) += JxW[qp]*(v_old*phi[i][qp]  // massmatrix term (1.theta)*dt*(U_old*grad_v_old)*phi[i][qp] + // convection term (1.theta)*dt*p_old*dphi[i][qp](1)  // pressure term on rhs (1.theta)*dt*(grad_v_old*dphi[i][qp]) + // diffusion term on rhs theta*dt*(U*grad_v)*phi[i][qp]); // Newton term ========================================================================================== Liang wrote: > Hi developers and users, > > I am looking at the nonlinear system of equations and want to borrow > some ideas from ex13, while the finite element equations in the > assemble function make me a little confused because I am not familiar > with the Stokes' equation, does any fluids expert might send me a > image format PDEs of that system? or give a link which is exact same > to the equations in ex13. I have looked at the Wikipedia but not make > sure which equations are belong to the case of ex13. Thanks a lot! > > Liang > 
From: Liang <goeasyon@gm...>  20100321 20:25:30

Hi developers and users, I have my nonlinear problem done, so please forget my previous questions on ex13 although the ex13 still doesn't convince me in the Newton term input. Thanks! Liang Liang wrote: > Hi developers and users, > > Seems I sent to wrong address last email, sorry for that. > > I tried to expand the tensor form of NS to the general PDEs as shown > in the attached picture, is this same to the governing equation that > ex13 used? please correct me if i am wrong. Is any book introduce > deriving the nonlinear PDEs to the element equations? nonlinear terms > in the equation bother me a lot, I don't know how to derive their weak > forms. > > Another question is about the constant \theta in the ex13, which is > equal to one in the equation, the description goes like: > > "Find out what the timestep size parameter is from the system, and the > value of theta for the theta method. We use implicit Euler (theta=1) > for this simulation even though it is only firstorder accurate in > time. " > > can I understand it as the coordinator for the CrankNicolson method, > theta =0.5 means the classic CrankNicolson, theta = 0/1 means pure > forward/backward Euler method? There are no \mu and \rho in the > assemble equation, so both parameters are evaluated one? > > Many thanks ! > > Liang > > > > ===============copy from ex13 ================== > > > > for (unsigned int i=0; i<n_u_dofs; i++) > { > Fu(i) += JxW[qp]*(u_old*phi[i][qp] >  // massmatrix term > (1.theta)*dt*(U_old*grad_u_old)*phi[i][qp] + > // convection term > (1.theta)*dt*p_old*dphi[i][qp](0)  > // pressure term on rhs > (1.theta)*dt*(grad_u_old*dphi[i][qp]) + > // diffusion term on rhs > theta*dt*(U*grad_u)*phi[i][qp]); > // Newton term > > Fv(i) += JxW[qp]*(v_old*phi[i][qp] >  // massmatrix term > (1.theta)*dt*(U_old*grad_v_old)*phi[i][qp] + > // convection term > (1.theta)*dt*p_old*dphi[i][qp](1)  > // pressure term on rhs > (1.theta)*dt*(grad_v_old*dphi[i][qp]) + > // diffusion term on rhs > theta*dt*(U*grad_v)*phi[i][qp]); > // Newton term > ========================================================================================== > > > > > > > Liang wrote: >> Hi developers and users, >> >> I am looking at the nonlinear system of equations and want to borrow >> some ideas from ex13, while the finite element equations in the >> assemble function make me a little confused because I am not familiar >> with the Stokes' equation, does any fluids expert might send me a >> image format PDEs of that system? or give a link which is exact same >> to the equations in ex13. I have looked at the Wikipedia but not make >> sure which equations are belong to the case of ex13. Thanks a lot! >> >> Liang >> > > >  > 
From: Liang <goeasyon@gm...>  20100322 14:01:19

Hi Developers and users, I have to say the libmesh is very excellent FEA package, I am extending my 2D case to the 3D, and want to try different element beside HEX8/27, is it any other 3D elements I can use? the second order should be fine. Thanks! Liang 
From: Kirk, Benjamin (JSCEG311) <benjamin.kirk1@na...>  20100322 14:56:18

> any other 3D elements I can use? the second order should be fine. > Thanks! Sure  tets, prisms, pyramids: http://libmesh.sourceforge.net/doxygen/classCell.php the only interface i am aware of in the library that can generate tet meshes is the TetGen interface. Ben 
From: John Peterson <peterson@cf...>  20100322 15:02:35

On Mon, Mar 22, 2010 at 9:56 AM, Kirk, Benjamin (JSCEG311) <benjamin.kirk1@...> wrote: >> any other 3D elements I can use? the second order should be fine. >> Thanks! > > Sure  tets, prisms, pyramids: > http://libmesh.sourceforge.net/doxygen/classCell.php > > the only interface i am aware of in the library that can generate tet meshes > is the TetGen interface. And you're much better off generating complicated grids with a GUI anyway. I'd recommend GMSH, they even have Win/Lin/Mac binaries available.  John 
From: Liang <goeasyon@gm...>  20100322 15:14:30

>> Sure  tets, prisms, pyramids: >> http://libmesh.sourceforge.net/doxygen/classCell.php >> >> the only interface i am aware of in the library that can generate tet meshes >> is the TetGen interface. >> > > And you're much better off generating complicated grids with a GUI > anyway. I'd recommend GMSH, they even have Win/Lin/Mac binaries > available. > > Ben, Thanks, this is a good page to learn the element scheme in libmesh, I will dig into it line by line. John, Thanks for the recommendation, I am going to download the GMSH and learn how to use it. Liang 