1 /* Copyright 2017 The TensorFlow Authors. All Rights Reserved. 2 3 Licensed under the Apache License, Version 2.0 (the "License"); 4 you may not use this file except in compliance with the License. 5 You may obtain a copy of the License at 6 7 http://www.apache.org/licenses/LICENSE-2.0 8 9 Unless required by applicable law or agreed to in writing, software 10 distributed under the License is distributed on an "AS IS" BASIS, 11 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 See the License for the specific language governing permissions and 13 limitations under the License. 14 ==============================================================================*/ 15 16 #ifndef TENSORFLOW_COMPILER_XLA_SERVICE_GPU_GEMM_THUNK_H_ 17 #define TENSORFLOW_COMPILER_XLA_SERVICE_GPU_GEMM_THUNK_H_ 18 19 #include "tensorflow/compiler/xla/service/buffer_assignment.h" 20 #include "tensorflow/compiler/xla/service/gpu/matmul_utils.h" 21 #include "tensorflow/compiler/xla/service/gpu/thunk.h" 22 #include "tensorflow/compiler/xla/status.h" 23 24 namespace xla { 25 namespace gpu { 26 27 // This is thread-compatible. 28 class GemmThunk : public Thunk { 29 public: 30 // Constructs a thunk that computes "output = (lhs <dot> rhs) * alpha" using 31 // BLAS gemm (alpha is stored in the instruction GemmBackendConfig). 32 GemmThunk(ThunkInfo thunk_info, GemmConfig config, 33 const BufferAllocation::Slice& lhs_buffer, 34 const BufferAllocation::Slice& rhs_buffer, 35 const BufferAllocation::Slice& output_buffer); 36 37 GemmThunk(const GemmThunk&) = delete; 38 GemmThunk& operator=(const GemmThunk&) = delete; 39 40 Status ExecuteOnStream(const ExecuteParams& params) override; 41 42 private: 43 const GemmConfig config_; 44 const BufferAllocation::Slice lhs_buffer_; 45 const BufferAllocation::Slice rhs_buffer_; 46 const BufferAllocation::Slice output_buffer_; 47 }; 48 49 } // namespace gpu 50 } // namespace xla 51 52 #endif // TENSORFLOW_COMPILER_XLA_SERVICE_GPU_GEMM_THUNK_H_ 53