dolfin team mailing list archive

[Nuno Lopes <question95433@xxxxxxxxxxxxxxxxxxxxx>]

Question #95433 on DOLFIN changed:
https://answers.launchpad.net/dolfin/+question/95433

Nuno Lopes gave more information on the question:
In my case part of the problem is definitely with uBLAS and Umfpack:
In 0.9.5 PETSc LU is much  faster than uBLAS LU (factorized solve)
-----------------------------------------------------------
Assemble cells                   |      0.011015       16.28  1478
Assemble exterior facets  |          0.01        0.02     2
Build sparsity                      |    1.3532e-05        0.02  1478
Delete sparsity                   |             0           0  1478
Init tensor                            |    6.7659e-06        0.01  1478
LU solver                            |      0.020638        6.15   298
X:DOLFWAVE:END          |         24.63       24.63     1
.........................................
Against the 83.71 s of uBLAS LU.

-- 
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.