On Wed, 7 Dec 2011, Xiangdong Liang wrote: > I think I did have 10 zeros entries. In my script, I have inserted > zeros by MatSetValue(A,i, (i+1)%N, 0.0, ADD_VALUES); Why doesn't that > option matter in my case? That option prevents PETSc from inserting > zeros, doesn't it? Thank you.
MatSetValue is a macro - perhaps its not processing the above properly. Try: PetscScalar zero=0.0 MatSetValue(A,i, (i+1)%N, zero, ADD_VALUES) And if its still not inzerting these zeros - check in a debugger. Satish
