| Prev | Next | ForJacSweep |
void ForJacSweep(
size_t npv,
size_t numvar,
const player<Base> *Rec,
size_t TaylorColDim,
const Base *Taylor,
Pack *ForJac
)
\[
F : B^n \rightarrow B^m
\]
ForJacSweep computes the sparsity pattern
for all the other variables.
Rec->TotNumVar().
i = 1 , \ldots , numvar
,
Taylor[i * TaylorColDim]
is the value of the variable with index i.
i = 1, \ldots , n
and
j = 0 , \ldots , npv
,
| field | Value |
ForJac[0 * npv + j] | the variable with index zero is not used |
Rec->GetOp(0) |
the operator with index zero must be a NonOp
|
ForJac[i * npv + j] | j-th subset of sparsity pattern for variable with index i |
Rec->GetOp(i) |
the operator with index i must be a InvOp
|
i = n+1, \ldots , numvar-1
,
j = 0 , \ldots , npv-1
,
and
k = n+1, \ldots ,
Rec->NumOp() - 1,
| field | Value |
ForJac[i * npv + j] | j-th set of sparsity pattern for variable with index i |
Rec->GetOp(i) |
any operator except for InvOp
|
i = 1, \ldots , n
and
j = 0 , \ldots , npv-1
,
Taylor[i * npv + j] is not modified.
i = m+1, \ldots , numvar-1
and
j = 0 , \ldots , npv-1
,
ForJac[i * npv + j] is set equal to the
j-th set of sparsity pattern for the variable with index i.