system command SparseSYM SOE

[seu]

In http://opensees.berkeley.edu/wiki/index ... rseSYM_SOE
it wrotes:

versions upto and including 2.2.0 use SparseSPD instead of SparseSYM as the option to the system command. For backward compatability this old option continues to work.

But my tcl model can work with "system SparseGeneral" while not "system SparseSYM", why? My opensees.exe version is 2.3.2.

The error message is as follow:
WARNING - ForceBeamColumn3d::update - failed to get compatible element forces &
deformations for element: 1(dW: << 1.#QNAN, dW0: 1.#QNAN)
WARNING - ForceBeamColumn3d::update - failed to get compatible element forces &
deformations for element: 2(dW: << 1.#QNAN, dW0: 1.#QNAN)
WARNING - ForceBeamColumn3d::update - failed to get compatible element forces &
deformations for element: 3(dW: << 1.#QNAN, dW0: 1.#QNAN)
...............

[vesna]

SparseSYM and SparseGeneral are different solvers. What one can solve the other one does not have to be able to solve it.

[seu]

http://opensees.berkeley.edu/wiki/index ... em_Command
As you told ,why there is not this SparseGeneral option here?
How about this statement:
"use SparseSPD instead of SparseSYM as the option to the system command"

[vesna]

It is there:
http://opensees.berkeley.edu/wiki/index.php/SuperLU_SOE

Where did you find the statement:
"use SparseSPD instead of SparseSYM as the option to the system command"

[seu]

http://opensees.berkeley.edu/wiki/index ... rseSYM_SOE
in the Notes paragraph.

[vesna]

Note says that SparseSYM and SparseSPD are the same solver. Before version 2.2.0 it was taught that it is symmetric positive definite so it was called sparseSPD, but later it was realized that it also solves negative definite problems so it was renamed to SparseSYM.

[seu]

Thanks ,you can always solve my problems.