The c_sysfn2() utility subroutine is exclusively used inside a FEloadsub user-written subroutine. The sole purpose of the c_sysfn2 utility subroutine is to register a scalar measure, for example, a position or a velocity component, needed for the desired computation. The c_sysfn2() subroutine has a C/C++ only implementation; there is no FORTRAN equivalent. The c_sysfn2() utility subroutine is supported only by Adams Solver C++.

The c_sysfn2() utility subroutine and associated FEloadsub user-written subroutine both implement a new paradigm. Users need to register the measures to be used by the subroutine only once when Adams Solver C++ calls the FEloadsub with the value of IFLAG equal to 1. When IFLAG is equal to 0 (normal evaluation call), all registered measures are evaluated by Adams Solver C++ and passed to the subroutine inside a data structure; see more details in FEloadsub.

There is strong similarity between c_sysfn2() and c_sysfnc() (and the FORTRAN version SYSFNC). The main difference is that c_sysfn2() does not evaluate the requested measure but just registers the measure. Adams Solver C++ will keep an associated data structure with all registered measures and compute the values of all the registered measures before calling the FEloadsub. The main similarity is in the use of parameters "fncnam", "ipar" and "nsize" as explained below.

FEloadsub exclusively.

The c_sysfnc2() subroutine registers a system-state value (such as displacement, velocities, non-dimensional arc length along a beam and so on.) to be used inside a FEloadsub user-written subroutine. Adams Solver C++ creates a list of all registered measures (the registration process takes place when IFLAG=1) and assigns an index to all registered measures. The index is 0-based and is based on the order in which the calls to c_sysfn2() are made. During normal evaluation (IFLAG=0), Adams Solver C++ computes the values of all registered measures and passes the computed values to the FEloadsub subroutine. See FEloadsub subroutine for more details and examples.

Although c_sysfn2() is similar to c_sysfnc() (or FORTRAN equivalent SYSFNC), the main difference is that users must call it only when IFLAG=1. This new architecture increases performance by eliminating continuous checks done on every call to SYSFNC; for example, during normal evaluation calls to user-written subroutines, every call to the utility subroutine SYSFNC requires a construction of the requested measure and a mandatory check to make sure the measure was registered. When using the c_sysfn2() utility, no constructions, no checks are needed during normal calls to the FEloadsub.

Calls to c_sysfn2() must be made only during the initialization stage (IFLAG=1).

In this example we present a code snippet where two measures (DX(2,8) and SD(3)) are registered.

In this complete example, four calls are made to c_sysfn2() in order to register the measures SD(2), SD(4), SD(5) and SD(6). The index is not used by the author of the subroutine.

This example is computational identical to the previous example; in this case the returned index from c_sysfn2() is used to make the code easier to understand.