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

BLAS and cuBLAS TPL interface for TRSM #532

Merged
merged 13 commits into from
Dec 20, 2019
+6,762 −0
Merged

BLAS and cuBLAS TPL interface for TRSM #532

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
fe510d2
Add BLAS and cuBLAS interface for TRSM
vqd8a Dec 13, 2019
1cdc8c2
Merge branch 'develop' into blas_trsm
vqd8a Dec 16, 2019
15302fe
Fix typo
vqd8a Dec 16, 2019
c3c89e3
Modify note for KK implementation
vqd8a Dec 16, 2019
96ab3d9
Use double quotes for KK headers and allow RHS with rank 2 only
vqd8a Dec 16, 2019
fe330f8
Add sequential fall-back implementation and change unit tests accordi…
vqd8a Dec 19, 2019
f2040de
Accidentally pushed this original file
vqd8a Dec 19, 2019
387640c
Accidentally pushed this original file
vqd8a Dec 19, 2019
85c76eb
Use KOKKOS_RESTRICT
vqd8a Dec 20, 2019
920b213
Merge branch 'blas_trsm' of github.com:kokkos/kokkos-kernels into bla…
vqd8a Dec 20, 2019
dda5321
Fix unused warning
vqd8a Dec 20, 2019
686b2bd
Merge branch 'develop' into blas_trsm
vqd8a Dec 20, 2019
f2394db
Add ETI files
vqd8a Dec 20, 2019
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
21 changes: 6 additions & 15 deletions src/blas/KokkosBlas3_trsm.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,9 +45,9 @@

/// \file KokkosBlas3_trsm.hpp

#include <KokkosKernels_Macros.hpp>
#include <KokkosBlas3_trsm_spec.hpp>
#include <KokkosKernels_helpers.hpp>
#include "KokkosKernels_Macros.hpp"
#include "KokkosBlas3_trsm_spec.hpp"
#include "KokkosKernels_helpers.hpp"
#include <sstream>
#include <type_traits>

Expand Down Expand Up @@ -92,8 +92,8 @@ trsm (const char side[],
"BViewType must be a Kokkos::View.");
static_assert (static_cast<int> (AViewType::rank) == 2,
"AViewType must have rank 2.");
static_assert (static_cast<int> (BViewType::rank) == 1 || static_cast<int> (BViewType::rank) == 2,
"BViewType must have either rank 1 or rank 2.");
static_assert (static_cast<int> (BViewType::rank) == 2,
"BViewType must have rank 2.");

// Check validity of indicator argument
bool valid_side = (side[0] == 'L' ) || (side[0] == 'l' )||
Expand Down Expand Up @@ -167,16 +167,7 @@ trsm (const char side[],
typename BViewType::device_type,
Kokkos::MemoryTraits<Kokkos::Unmanaged> >;

AVT A_i = A;

if (BViewType::rank == 1) {
auto B_i = BVT(B.data(), B.extent(0), 1);
KokkosBlas::Impl::TRSM<AVT, BVT>::trsm (side, uplo, trans, diag, alpha, A_i, B_i);
}
else { //BViewType::rank == 2
auto B_i = BVT(B.data(), B.extent(0), B.extent(1));
KokkosBlas::Impl::TRSM<AVT, BVT>::trsm (side, uplo, trans, diag, alpha, A_i, B_i);
}
KokkosBlas::Impl::TRSM<AVT, BVT>::trsm (side, uplo, trans, diag, alpha, A, B);
}

} // namespace KokkosBlas
Expand Down
2 changes: 1 addition & 1 deletion src/blas/impl/KokkosBlas3_trsm_impl.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,7 +54,7 @@

namespace KokkosBlas {
namespace Impl {
//TODO: Add the implementation of KokkosBlas::trsm later
//Note: Add an implementation of KokkosBlas::trsm in a next PR.
}
}
#endif // KOKKOSBLAS3_TRSM_IMPL_HPP_
2 changes: 1 addition & 1 deletion src/blas/impl/KokkosBlas3_trsm_spec.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -130,7 +130,7 @@ struct TRSM {
std::ostringstream os;
os << "KokkosBlas::trsm currently supports only TPL interface."
<< "Please enable Host BLAS or cuBLAS."
<< "KokkosKernels implmentation will be added in future release.";
<< "KokkosKernels implementation will be added in future release.";
mhoemmen marked this conversation as resolved.
Show resolved Hide resolved
Kokkos::Impl::throw_runtime_exception (os.str ());
}
};
Expand Down