On May 9, 2011, at 11:37 PM, Blaise Bourdin wrote: > Barry, > >>>> is there any way I can reproduce the bugs without going through all that >>>> crap? >>> >>> Let me see if I can build a Mesh from an exo file in C, and save it in a >>> binary file, once MeshLoad is fixed, we can replicate the other bugs from >>> there. >> >> >> Ok, let's try that. > > I am attaching 2 simple programs and a datafile. > Beam3B_s.dat was obtained using either the C or fortran program using 1 mpi > task, reading an exo file and writing it to a file using DMView. I obtain a > text file instead of a binary file, but I can read it back onto one cpu (use > -mode 1). Distributed IO is broken. > > Once we can save and load proper binary files, I will generate binary mesh > files too. > > Blaise >
Ok, after removing an extra PetscFinalize() from TestMeshf2.F90 I was able to run it with mode 1 successfully and it printed some ASCII to the screen with no memory problems. Now what should I do with these codes to generate one of the PETSc errors you reported? (Or was it suppose to crash?) Thanks Barry > > <Beam3B_s.dat><Beam3B_s.dat.info><TestMesh.c><TestMeshf2.F90> > -- > Department of Mathematics and Center for Computation & Technology > Louisiana State University, Baton Rouge, LA 70803, USA > Tel. +1 (225) 578 1612, Fax +1 (225) 578 4276 > http://www.math.lsu.edu/~bourdin > > > > > > >