Home | History | Annotate | Download | only in gpu
      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_SEQUENTIAL_THUNK_H_
     17 #define TENSORFLOW_COMPILER_XLA_SERVICE_GPU_SEQUENTIAL_THUNK_H_
     18 
     19 #include <vector>
     20 
     21 #include "tensorflow/compiler/xla/service/gpu/buffer_allocations.h"
     22 #include "tensorflow/compiler/xla/service/gpu/thunk.h"
     23 #include "tensorflow/compiler/xla/service/hlo_instruction.h"
     24 #include "tensorflow/core/platform/stream_executor_no_cuda.h"
     25 
     26 namespace xla {
     27 namespace gpu {
     28 
     29 // A thunk that wraps a list of sub-thunks. Executing this thunk executes all
     30 // the sub-thunks sequentially. This is useful to implement instructions that
     31 // require multiple kernel launches or library calls.
     32 class SequentialThunk : public Thunk {
     33  public:
     34   SequentialThunk(std::vector<std::unique_ptr<Thunk>>&& thunks,
     35                   const HloInstruction* hlo);
     36   SequentialThunk(const SequentialThunk&) = delete;
     37   SequentialThunk& operator=(const SequentialThunk&) = delete;
     38 
     39   const std::vector<std::unique_ptr<Thunk>>& thunks() const { return thunks_; }
     40 
     41   tensorflow::Status Initialize(const GpuExecutable& executable) override;
     42   tensorflow::Status ExecuteOnStream(
     43       const BufferAllocations& buffer_allocations,
     44       perftools::gputools::Stream* stream) override;
     45 
     46  private:
     47   // The list of sub-thunks.
     48   std::vector<std::unique_ptr<Thunk>> thunks_;
     49 };
     50 
     51 }  // namespace gpu
     52 }  // namespace xla
     53 
     54 #endif  // TENSORFLOW_COMPILER_XLA_SERVICE_GPU_SEQUENTIAL_THUNK_H_
     55