I know some people had a similar issue with non-cleaned cache. If the problem persists, I'll fix it as soon as I have a failing test case. Thanks -- Fabio 2016-05-10 12:11 GMT+02:00 Mitchell, Lawrence < lawrence.mitchell@imperial.ac.uk>:
On 10 May 2016, at 11:02, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
I did do the updates and all went through well it seems, but both with the codes of Fryderyk and Will matters crash at the solve stage; I also attache d F's code he just send me. I run on an macbook and they run under linux. The code runs for Fryderyk, he said.
mtob$ python ns_plane_poiseuille.py ... Traceback (most recent call last): File "ns_plane_poiseuille.py", line 104, in <module> solve(F_u == 0, u, bcs = [bc0,bc1,bc2],solver_parameters={'pc_type': 'lu', 'pc_factor_mat_solver_package': 'mumps'} ) # File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/solving.py", line 120, in solve _solve_varproblem(*args, **kwargs) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/solving.py", line 164, in _solve_varproblem solver.solve() File "<decorator-gen-295>", line 2, in solve File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/pyop2/profiling.py", line 203, in wrapper return f(*args, **kwargs) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/variational_solver.py", line 190, in solve solving_utils.check_snes_convergence(self.snes) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 62, in check_snes_convergence %s""" % (snes.getIterationNumber(), msg)) RuntimeError: Nonlinear solve failed to converge after 0 nonlinear iterations. Reason: Inner linear solve failed to converge after 0 iterations with reason: unknown reason (petsc4py enum incomplete?
This looks like the jacobian is singular, please follow the advice here: http://firedrakeproject.org/solving-interface.html#debugging-convergence-fai...
To try and get an idea as to what is going on.
I get: Linear solve did not converge due to DIVERGED_PCSETUP_FAILED iterations 0 PCSETUP_FAILED due to FACTOR_NUMERIC_ZEROPIVOT
Similar problem for Will's code: ast_handler.plan_cpu(self._opts) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/plan.py", line 184, in plan_cpu _generate_cpu_code(self, kernel, **params) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/plan.py", line 110, in _generate_cpu_code loop_opt.eliminate_zeros() File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/optimizer.py", line 163, in eliminate_zeros self.nz_syms = zls.reschedule(self.header) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/scheduler.py", line 844, in reschedule self.exprs.update(elf.fission(stmt, expr_info)) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/scheduler.py", line 463, in fission split, splittable = self.cutter.cut(*splittable) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/scheduler.py", line 300, in cut index = expr_info.parent.children.index(node) ValueError: <coffee.base.Incr object at 0x10e4186d0> is not in list
This may be a bug in coffee, but it runs for Will? Please try running firedrake-clean and then try again if so. Otherwise, please provide a failing test case and report it as a github issue ( https://github.com/firedrakeproject/firedrake/issues)
Thanks,
Lawrence
Could you run firedrake-update, then firedrake-clean, and re-run the failing test case ? We've just pushed a fix Thanks -- Fabio 2016-05-10 11:23 GMT+01:00 Fabio Luporini <f.luporini12@imperial.ac.uk>:
I know some people had a similar issue with non-cleaned cache. If the problem persists, I'll fix it as soon as I have a failing test case. Thanks
-- Fabio
2016-05-10 12:11 GMT+02:00 Mitchell, Lawrence < lawrence.mitchell@imperial.ac.uk>:
On 10 May 2016, at 11:02, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
I did do the updates and all went through well it seems, but both with the codes of Fryderyk and Will matters crash at the solve stage; I also attache d F's code he just send me. I run on an macbook and they run under linux. The code runs for Fryderyk, he said.
mtob$ python ns_plane_poiseuille.py ... Traceback (most recent call last): File "ns_plane_poiseuille.py", line 104, in <module> solve(F_u == 0, u, bcs = [bc0,bc1,bc2],solver_parameters={'pc_type': 'lu', 'pc_factor_mat_solver_package': 'mumps'} ) # File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/solving.py", line 120, in solve _solve_varproblem(*args, **kwargs) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/solving.py", line 164, in _solve_varproblem solver.solve() File "<decorator-gen-295>", line 2, in solve File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/pyop2/profiling.py", line 203, in wrapper return f(*args, **kwargs) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/variational_solver.py", line 190, in solve solving_utils.check_snes_convergence(self.snes) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 62, in check_snes_convergence %s""" % (snes.getIterationNumber(), msg)) RuntimeError: Nonlinear solve failed to converge after 0 nonlinear iterations. Reason: Inner linear solve failed to converge after 0 iterations with reason: unknown reason (petsc4py enum incomplete?
This looks like the jacobian is singular, please follow the advice here: http://firedrakeproject.org/solving-interface.html#debugging-convergence-fai...
To try and get an idea as to what is going on.
I get: Linear solve did not converge due to DIVERGED_PCSETUP_FAILED iterations 0 PCSETUP_FAILED due to FACTOR_NUMERIC_ZEROPIVOT
Similar problem for Will's code: ast_handler.plan_cpu(self._opts) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/plan.py", line 184, in plan_cpu _generate_cpu_code(self, kernel, **params) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/plan.py", line 110, in _generate_cpu_code loop_opt.eliminate_zeros() File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/optimizer.py", line 163, in eliminate_zeros self.nz_syms = zls.reschedule(self.header) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/scheduler.py", line 844, in reschedule self.exprs.update(elf.fission(stmt, expr_info)) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/scheduler.py", line 463, in fission split, splittable = self.cutter.cut(*splittable) File "/Users/amtob/werk/firedrake/lib/python2.7/site-packages/coffee/scheduler.py", line 300, in cut index = expr_info.parent.children.index(node) ValueError: <coffee.base.Incr object at 0x10e4186d0> is not in list
This may be a bug in coffee, but it runs for Will? Please try running firedrake-clean and then try again if so. Otherwise, please provide a failing test case and report it as a github issue ( https://github.com/firedrakeproject/firedrake/issues)
Thanks,
Lawrence
participants (1)
-
Fabio Luporini