//@HEADER // ************************************************************************ // // Kokkos v. 4.0 // Copyright (2022) National Technology & Engineering // Solutions of Sandia, LLC (NTESS). // // Under the terms of Contract DE-NA0003525 with NTESS, // the U.S. Government retains certain rights in this software. // // Part of Kokkos, under the Apache License v2.0 with LLVM Exceptions. // See https://kokkos.org/LICENSE for license information. // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception // //@HEADER #ifndef KOKKOSBATCHED_GEMM_TEAM_INTERNAL_HPP #define KOKKOSBATCHED_GEMM_TEAM_INTERNAL_HPP /// \author Kyungjoo Kim (kyukim@sandia.gov) #include "KokkosBatched_Util.hpp" #include "KokkosKernels_ExecSpaceUtils.hpp" #include "KokkosBlas1_set_impl.hpp" #include "KokkosBlas1_team_scal_impl.hpp" #include "KokkosBatched_InnerGemmFixC_Serial_Impl.hpp" namespace KokkosBatched { /// /// Team Internal Impl /// ==================== template struct TeamGemmInternal { template KOKKOS_INLINE_FUNCTION static int invoke(const MemberType &member, const int m, const int n, const int k, const ScalarType alpha, const ValueType *KOKKOS_RESTRICT A, const int as0, const int as1, const ValueType *KOKKOS_RESTRICT B, const int bs0, const int bs1, const ScalarType beta, /**/ ValueType *KOKKOS_RESTRICT C, const int cs0, const int cs1); }; template <> template KOKKOS_INLINE_FUNCTION int TeamGemmInternal::invoke( const MemberType &member, const int m, const int n, const int k, const ScalarType alpha, const ValueType *KOKKOS_RESTRICT A, const int as0, const int as1, const ValueType *KOKKOS_RESTRICT B, const int bs0, const int bs1, const ScalarType beta, /**/ ValueType *KOKKOS_RESTRICT C, const int cs0, const int cs1) { // C = beta C + alpha A B // C (m x n), A(m x k), B(k x n) const ScalarType one(1.0), zero(0.0); if (beta == zero) KokkosBlas::Impl::TeamSetInternal::invoke(member, m, n, zero, C, cs0, cs1); else if (beta != one) KokkosBlas::Impl::TeamScaleInternal::invoke(member, m, n, beta, C, cs0, cs1); if (alpha != ScalarType(0.0)) { if (m <= 0 || n <= 0 || k <= 0) return 0; if (beta != one) member.team_barrier(); Kokkos::parallel_for(Kokkos::TeamThreadRange(member, 0, m * n), [&](const int &ij) { // assume layout right for batched computation const int i = ij / n, j = ij % n; const ValueType *KOKKOS_RESTRICT pA = A + i * as0, *KOKKOS_RESTRICT pB = B + j * bs1; ValueType c = ValueType(0); for (int p = 0; p < k; ++p) c += pA[p * as1] * pB[p * bs0]; C[i * cs0 + j * cs1] += alpha * c; }); } return 0; } template <> template KOKKOS_INLINE_FUNCTION int TeamGemmInternal::invoke( const MemberType &member, const int m, const int n, const int k, const ScalarType alpha, const ValueType *KOKKOS_RESTRICT A, const int as0, const int as1, const ValueType *KOKKOS_RESTRICT B, const int bs0, const int bs1, const ScalarType beta, /**/ ValueType *KOKKOS_RESTRICT C, const int cs0, const int cs1) { // C = beta C + alpha A B // C (m x n), A(m x k), B(k x n) constexpr int mbAlgo = Algo::Gemm::Blocked::mb(); constexpr int nbAlgo = Algo::Gemm::Blocked::mb(); const ScalarType one(1.0), zero(0.0); if (beta == zero) KokkosBlas::Impl::TeamSetInternal::invoke(member, m, n, zero, C, cs0, cs1); else if (beta != one) KokkosBlas::Impl::TeamScaleInternal::invoke(member, m, n, beta, C, cs0, cs1); if (alpha != ScalarType(0.0)) { if (m <= 0 || n <= 0 || k <= 0) return 0; if (beta != one) member.team_barrier(); /// /// GPU case: team size is large and blocksize (mb,nb) is small InnerGemmFixC inner(as0, as1, bs0, bs1, cs0, cs1); auto gemm = [&](const int ib, const int jb, const int pb, const ValueType *KOKKOS_RESTRICT AA, const ValueType *KOKKOS_RESTRICT BB, /**/ ValueType *KOKKOS_RESTRICT CC) { // Made this non-const in order to WORKAROUND issue #349 int mb = mbAlgo, mp = (ib % mb), mq = (ib / mb) + (mp > 0), nb = nbAlgo, np = (jb % nb), nq = (jb / nb) + (np > 0); // square tiling Kokkos::parallel_for(Kokkos::TeamThreadRange(member, mq * nq), [&](const int &ij) { int i, j; // note: the condition is constexpr if (KokkosKernels::Impl::is_gpu_exec_space_v) { i = ij % mq * mb; j = ij / mq * nb; } else { i = ij / nq * mb; j = ij % nq * nb; } inner.serial_invoke(alpha, AA + i * as0, BB + j * bs1, (i + mb) > ib ? mp : mb, (j + nb) > jb ? np : nb, pb, CC + i * cs0 + j * cs1); }); }; const bool is_small = true; //(m*n*k <= 64*64*64); if (is_small) { gemm(m, n, k, A, B, C); } else { // // cache blocking // const int // nc = nb*10, kc = mb*4, mc = mb*4; // for (int jj=0;jj