Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Incomplete LU factors

Please login with a confirmed email address before reporting spam

I see this message a lot in my log file. What does it mean?
"Incomplete LU factors 511842 nonzeros (original matrix has 381953)"

Thanks,
Sirisha

5 Replies Last Post 2011年11月2日 GMT-4 18:16
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年11月2日 GMT-4 02:31
Hi

its linked to the matrix decomposition (LU factoring) I suppose it could come from missing BCs or incomplete BC values. You need to give enough boundary conditions to enforece a single solution to your ODEs. COMSOl cannot handle multiple solutions without your help

--
Good luck
Ivar
Hi its linked to the matrix decomposition (LU factoring) I suppose it could come from missing BCs or incomplete BC values. You need to give enough boundary conditions to enforece a single solution to your ODEs. COMSOl cannot handle multiple solutions without your help -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年11月2日 GMT-4 10:52
But the same error disappears when I use a direct-fullycoupled-solver instead of a segregated solver. In the segregated solver, I am using a combination of direct and iterative solvers (one for fluid variables and the other for solid variables).

So if it is a BC issue then I don't understand why the error does not occur for the direct-fullycoupled-solver.

Sirisha
But the same error disappears when I use a direct-fullycoupled-solver instead of a segregated solver. In the segregated solver, I am using a combination of direct and iterative solvers (one for fluid variables and the other for solid variables). So if it is a BC issue then I don't understand why the error does not occur for the direct-fullycoupled-solver. Sirisha

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年11月2日 GMT-4 11:13
Hi

then it must be linked with the way you define your segregated solver, the order of the variables must be such that they depend on each other in a logical way. A regrouping issue I believe.

Why not use the direct solver, its mostly quicker then

--
Good luck
Ivar
Hi then it must be linked with the way you define your segregated solver, the order of the variables must be such that they depend on each other in a logical way. A regrouping issue I believe. Why not use the direct solver, its mostly quicker then -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年11月2日 GMT-4 11:31
Well........using a direct solver is ok when I am doing 2D but not with 3D due to the memory issues.
I will try to change the order of the variables and see how things go.

Thanks,
Sirisha
Well........using a direct solver is ok when I am doing 2D but not with 3D due to the memory issues. I will try to change the order of the variables and see how things go. Thanks, Sirisha

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年11月2日 GMT-4 18:16
I have not upgraded to 4.2a yet but looks like there is information about this warning in the 4.2a reference guide. I contacted the customer support about my question and here is their response:

" That appears as a warning message and is not an error by itself. The
message tells details about the preconditioner of a matrix. This warning
message should not be of concern to you unless there are convergence
issues. You can read more about it on page number 528 of the Reference
Guide"

Thanks,
Sirisha

I have not upgraded to 4.2a yet but looks like there is information about this warning in the 4.2a reference guide. I contacted the customer support about my question and here is their response: " That appears as a warning message and is not an error by itself. The message tells details about the preconditioner of a matrix. This warning message should not be of concern to you unless there are convergence issues. You can read more about it on page number 528 of the Reference Guide" Thanks, Sirisha

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.