Data Type

Integer

Default Value

1

Description

LOWASM specifies the first equation in the matrix which has a term that changed. By properly locating the equations containing nonlinearities at the bottom of the matrix, you may save a significant amount of computation.
LOWASM
FMS will round the value of LOWASM down to start on a segment boundary. This prevents records from containing a mix of factored and unfactored data.

FMS begins assembling or factoring the new matrix at the segment containing LOWASM. For Profile or Block format matrices, this is the row number in the lower triangle [AL] and column number in the upper triangle [AU]. For Slab format matrices, this is the column number in [AL] and [AU]. As the new matrix factors [L] and [U] are computed, they are overwritten on the files containing the previous factors.

One benefit of this feature is that you do not need to know the restart point in advance when the matrix [A] is factored the first time. You simply determine LOWASM as you are processing the solution {X} and FMS automatically skips the appropriate calculations on the next solution.

For example, in a structural finite element analysis, this feature would be used as follows:

  1. Initialize your value of LOWASM to 1 greater than the number of equations.
  2. Loop over the elements. Use the displacements {X} from the previous solution to compute new values of strains and stresses.
  3. If nonlinear effects are observed due to large deformations or material nonlinearities, set LOWASM=MIN0(LOWASM,ILOW), where ILOW is the lowest equation in the element.
  4. CALL FMSIST('LOWASM',LOWASM)
    to pass your value of LOWASM to FMS.
This parameter may also be used to recover a factoring process that terminates prematurely. During the initial factoring, set the FMS Parameter IREST to write the restart data. On the second (recovery) run, set this FMS parameter LOWASM to the highest factored segment (or slab) printed by FMS during the first run.