Precision of Floating Point Arithmetic in Solver |
  
|
Precision of Floating Point Arithmetic in Solver |
|
The option to select the precision of floating point arithmetic is located in the Analysis Options dialog, which can be opened by clicking Analysis > Analysis Options.
Double precision solver:
•standard solver similar to many analysis programs in the market
•enough for most structures
Double precision Sparse solver:
•fastest, but lacks informative messages when something goes wrong during a solution
•can only be used for static analysis
•useful to solve extremely large structural models
•uses an out-of-core approach to minimize the requirement of computer memory
Quad precision Skyline solver:
•provides an invaluable alternative for some large or numerically sensitive models where the 64-bit floating point (double precision) solver may fail
•extremely stable and accurate, but relatively slow
•the recommended solver if the model contains rigid diaphragms to avoid numerical difficulties.
