Thread Previous • Date Previous • Date Next • Thread Next |
Alessio Quaglino wrote:
Alessio Quaglino wrote:With "the mesh remains constant" is meant that the vertices cannot move or that there should not be any connectivity change?No connectivity changes.I see that in void Assembler::assemble(GenericTensor& A, const Form& form, bool reset_tensor) { assemble(A, form.form(), form.coefficients(), 0, 0, 0); } reset_tensor is not used. Is this wanted? The initGlobalTensor takes 90% of my runtime (I think it wasn't like this before the last release).
That's exactly why it's there. If the matrix has already been initialised, then there is not need to re-initialise. This will typically be useful for nonlinear problems.
If initialising the matrix is requiring 90% of your runtime, you're probably doing something strange. Initialising the matrix does now take more time, with the trade-off that assembly is now much faster and less memory is required.
Garth
AlessioGarthAlessioOne or more new changesets pushed to the primary dolfin repository. A short summary of the last three changesets is included below. changeset: 3358:fe232f61370532bd51b0f77df3fdd6aeaf5aa82c tag: tip user: "Anders Logg <logg at simula.no>" date: Tue Aug 28 14:50:21 2007 +0200 files: src/kernel/fem/Assembler.cpp src/kernel/fem/assemble.cpp src/kernel/fem/dolfin/Assembler.h src/kernel/pde/LinearPDE.cpp src/kernel/pde/NonlinearPDE.cpp description: Add mesh to constructor of Assembler. Reuse of dof maps requires that the mesh remains constant. changeset: 3357:4ad8f4efd1d6c87e9d6dc82d85e2d8f3e4facff0 user: "Anders Logg <logg at simula.no>" date: Tue Aug 28 14:26:43 2007 +0200 files: src/kernel/function/UserFunction.cpp src/kernel/function/dolfin/Function.h src/kernel/function/dolfin/UserFunction.h description: Make rank() and dim() virtual (overloadable by user). changeset: 3356:24c98268945163d11ab3b45f08f6015c25b4546c user: "Anders Logg <logg at simula.no>" date: Tue Aug 28 13:38:22 2007 +0200 files: src/demo/pde/poisson/python/demo.py description: test ---------------------------------------------------------------------- For more details, visit http://www.fenics.org/hg/dolfin_______________________________________________ DOLFIN-dev mailing list DOLFIN-dev at fenics.org http://www.fenics.org/mailman/listinfo/dolfin-dev_______________________________________________ DOLFIN-dev mailing list DOLFIN-dev@xxxxxxxxxx http://www.fenics.org/mailman/listinfo/dolfin-dev
Thread Previous • Date Previous • Date Next • Thread Next |