Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Thyra DefaultMultipliedLinearOp: Caching of intermediate vectors #13738

Merged
merged 1 commit into from
Jan 22, 2025
Merged

Thyra DefaultMultipliedLinearOp: Caching of intermediate vectors #13738

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions .../core/src/support/operator_vector/client_support/Thyra_DefaultMultipliedLinearOp_decl.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -212,6 +212,8 @@ class DefaultMultipliedLinearOp : virtual public MultipliedLinearOpBase<Scalar>
*/
bool opSupportedImpl(EOpTransp M_trans) const;

void allocateVecs(const Ordinal dim) const;

/** \brief . */
void applyImpl(
const EOpTransp M_trans,
Expand All @@ -228,6 +230,7 @@ class DefaultMultipliedLinearOp : virtual public MultipliedLinearOpBase<Scalar>
private:

Array<Teuchos::ConstNonconstObjectContainer<LinearOpBase<Scalar> > > Ops_;
mutable std::vector<Teuchos::RCP<MultiVectorBase<Scalar> > > T_k_;

inline void assertInitialized() const;
inline std::string getClassName() const;
Expand Down
22 changes: 20 additions & 2 deletions ...a/core/src/support/operator_vector/client_support/Thyra_DefaultMultipliedLinearOp_def.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -236,6 +236,23 @@ bool DefaultMultipliedLinearOp<Scalar>::opSupportedImpl(EOpTransp M_trans) const
// ToDo: Cache these?
}

template<class Scalar>
void DefaultMultipliedLinearOp<Scalar>::allocateVecs(const Ordinal dim) const {
const int nOps = Ops_.size();
if ((T_k_.size() != Teuchos::as<size_t>(nOps+1)) || ((nOps > 0) && (T_k_[0]->domain()->dim() != dim))) {
// op[0]->range
// op[0]->domain == op[1]->range
// ...
// op[nOps-2]->domain == op[nOps-1]->range
// op[nOps-1]->domain
T_k_.resize(0);
for( int k = 0; k < nOps; ++k ) {
T_k_.push_back(createMembers(getOp(k)->range(), dim));
}
T_k_.push_back(createMembers(getOp(nOps-1)->domain(), dim));
}
}


template<class Scalar>
void DefaultMultipliedLinearOp<Scalar>::applyImpl(
Expand All @@ -255,6 +272,7 @@ void DefaultMultipliedLinearOp<Scalar>::applyImpl(
#endif // TEUCHOS_DEBUG
const int nOps = Ops_.size();
const Ordinal m = X.domain()->dim();
allocateVecs(m);
if( real_trans(M_trans)==NOTRANS ) {
//
// Y = alpha * M * X + beta*Y
Expand All @@ -265,7 +283,7 @@ void DefaultMultipliedLinearOp<Scalar>::applyImpl(
for( int k = nOps-1; k >= 0; --k ) {
RCP<MultiVectorBase<Scalar> > Y_k;
RCP<const MultiVectorBase<Scalar> > X_k;
if(k==0) Y_k = rcpFromPtr(Y); else Y_k = T_k = createMembers(getOp(k)->range(), m);
if(k==0) Y_k = rcpFromPtr(Y); else Y_k = T_k = T_k_[k];
if(k==nOps-1) X_k = rcpFromRef(X); else X_k = T_kp1;
if( k > 0 )
Thyra::apply(*getOp(k), M_trans, *X_k, Y_k.ptr());
Expand All @@ -284,7 +302,7 @@ void DefaultMultipliedLinearOp<Scalar>::applyImpl(
for( int k = 0; k <= nOps-1; ++k ) {
RCP<MultiVectorBase<Scalar> > Y_k;
RCP<const MultiVectorBase<Scalar> > X_k;
if(k==nOps-1) Y_k = rcpFromPtr(Y); else Y_k = T_k = createMembers(getOp(k)->domain(), m);
if(k==nOps-1) Y_k = rcpFromPtr(Y); else Y_k = T_k = T_k_[k+1];
if(k==0) X_k = rcpFromRef(X); else X_k = T_km1;
if( k < nOps-1 )
Thyra::apply(*getOp(k), M_trans, *X_k, Y_k.ptr());
Expand Down
Loading