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.

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

Jgfouca/block spiluk fixes #2172

Merged
merged 11 commits into from
Apr 9, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
46 changes: 44 additions & 2 deletions batched/dense/impl/KokkosBatched_Trsm_Team_Impl.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -99,6 +99,48 @@ struct TeamTrsm<MemberType, Side::Right, Uplo::Upper, Trans::NoTranspose,
}
};

///
/// R/L/NT
///
/// B := (alpha*B) inv(tril(A))
/// A(n x n), B(m x n)

template <typename MemberType, typename ArgDiag>
struct TeamTrsm<MemberType, Side::Right, Uplo::Lower, Trans::NoTranspose,
ArgDiag, Algo::Trsm::Unblocked> {
template <typename ScalarType, typename AViewType, typename BViewType>
KOKKOS_INLINE_FUNCTION static int invoke(const MemberType &member,
const ScalarType alpha,
const AViewType &A,
const BViewType &B) {
return TeamTrsmInternalLeftUpper<Algo::Trsm::Unblocked>::invoke(
member, ArgDiag::use_unit_diag, B.extent(1), B.extent(0), alpha,
A.data(), A.stride_1(), A.stride_0(), B.data(), B.stride_1(),
B.stride_0());
}
};

template <typename MemberType, typename ArgDiag>
struct TeamTrsm<MemberType, Side::Right, Uplo::Lower, Trans::NoTranspose,
ArgDiag, Algo::Trsm::Blocked> {
template <typename ScalarType, typename AViewType, typename BViewType>
KOKKOS_INLINE_FUNCTION static int invoke(const MemberType &member,
const ScalarType alpha,
const AViewType &A,
const BViewType &B) {
return TeamTrsmInternalLeftUpper<Algo::Trsm::Blocked>::invoke(
member, ArgDiag::use_unit_diag, B.extent(1), B.extent(0), alpha,
A.data(), A.stride_1(), A.stride_0(), B.data(), B.stride_1(),
B.stride_0());
}
};

///
/// R/U/T
///
/// B := (alpha*B) inv(triu(A))
/// A(n x n), B(m x n)

template <typename MemberType, typename ArgDiag>
struct TeamTrsm<MemberType, Side::Right, Uplo::Upper, Trans::Transpose, ArgDiag,
Algo::Trsm::Unblocked> {
Expand All @@ -107,7 +149,7 @@ struct TeamTrsm<MemberType, Side::Right, Uplo::Upper, Trans::Transpose, ArgDiag,
const ScalarType alpha,
const AViewType &A,
const BViewType &B) {
return TeamTrsmInternalLeftLower<Algo::Trsm::Unblocked>::invoke(
return TeamTrsmInternalLeftUpper<Algo::Trsm::Unblocked>::invoke(
member, ArgDiag::use_unit_diag, B.extent(1), B.extent(0), alpha,
A.data(), A.stride_0(), A.stride_1(), B.data(), B.stride_1(),
B.stride_0());
Expand All @@ -122,7 +164,7 @@ struct TeamTrsm<MemberType, Side::Right, Uplo::Upper, Trans::Transpose, ArgDiag,
const ScalarType alpha,
const AViewType &A,
const BViewType &B) {
return TeamTrsmInternalLeftLower<Algo::Trsm::Blocked>::invoke(
return TeamTrsmInternalLeftUpper<Algo::Trsm::Blocked>::invoke(
member, ArgDiag::use_unit_diag, B.extent(1), B.extent(0), alpha,
A.data(), A.stride_0(), A.stride_1(), B.data(), B.stride_1(),
B.stride_0());
Expand Down
Loading