Cannot access raw-matrix for mixed function spaces
Issue #497
invalid
Hello,
if i try to access A.array() after the assembly within the demo_stokes-iterative.py example, a memory error is raised:
---------------------------------------------------------------------------
MemoryError Traceback (most recent call last)
<ipython-input-5-dc6137c9900a> in <module>()
----> 1 A.array()
/usr/lib/python2.7/dist-packages/dolfin/cpp/la.pyc in array(self)
571 from numpy import zeros
572 m_range = self.local_range(0);
--> 573 A = zeros((m_range[1] - m_range[0], self.size(1)))
574 for i, row in enumerate(range(*m_range)):
575 column, values = self.getrow(row)
MemoryError:
doing the same within demo_poisson.py gives a correct output, so i could be a problem with the mixed function space.
greeting FloB
Comments (2)
-
-
- changed status to invalid
A.array()
is too big to fit in memory. If you change the problem size toUnitCubeMesh(2, 2, 2)
for example, it will work. - Log in to comment
I have a suspicion that you get really run out of memory when asking for
A.array()
which is dense.