|
| 1 | +// Copyright 2021, NVIDIA CORPORATION & AFFILIATES. All rights reserved. |
| 2 | +// |
| 3 | +// Redistribution and use in source and binary forms, with or without |
| 4 | +// modification, are permitted provided that the following conditions |
| 5 | +// are met: |
| 6 | +// * Redistributions of source code must retain the above copyright |
| 7 | +// notice, this list of conditions and the following disclaimer. |
| 8 | +// * Redistributions in binary form must reproduce the above copyright |
| 9 | +// notice, this list of conditions and the following disclaimer in the |
| 10 | +// documentation and/or other materials provided with the distribution. |
| 11 | +// * Neither the name of NVIDIA CORPORATION nor the names of its |
| 12 | +// contributors may be used to endorse or promote products derived |
| 13 | +// from this software without specific prior written permission. |
| 14 | +// |
| 15 | +// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY |
| 16 | +// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 17 | +// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR |
| 18 | +// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR |
| 19 | +// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, |
| 20 | +// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
| 21 | +// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR |
| 22 | +// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY |
| 23 | +// OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 24 | +// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| 25 | +// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 26 | + |
| 27 | +#include "infer_request.h" |
| 28 | + |
| 29 | +#ifdef TRITON_PB_STUB |
| 30 | +#include "infer_response.h" |
| 31 | +#include "pb_stub.h" |
| 32 | +#endif |
| 33 | + |
| 34 | +namespace triton { namespace backend { namespace python { |
| 35 | + |
| 36 | +InferRequest::InferRequest( |
| 37 | + const std::string& request_id, uint64_t correlation_id, |
| 38 | + const std::vector<std::shared_ptr<PbTensor>>& inputs, |
| 39 | + const std::vector<std::string>& requested_output_names, |
| 40 | + const std::string& model_name, const int64_t model_version) |
| 41 | + : request_id_(request_id), correlation_id_(correlation_id), inputs_(inputs), |
| 42 | + requested_output_names_(requested_output_names), model_name_(model_name), |
| 43 | + model_version_(model_version) |
| 44 | +{ |
| 45 | +} |
| 46 | + |
| 47 | +const std::vector<std::shared_ptr<PbTensor>>& |
| 48 | +InferRequest::Inputs() |
| 49 | +{ |
| 50 | + return inputs_; |
| 51 | +} |
| 52 | + |
| 53 | +const std::string& |
| 54 | +InferRequest::RequestId() |
| 55 | +{ |
| 56 | + return request_id_; |
| 57 | +} |
| 58 | + |
| 59 | +uint64_t |
| 60 | +InferRequest::CorrelationId() |
| 61 | +{ |
| 62 | + return correlation_id_; |
| 63 | +} |
| 64 | + |
| 65 | +const std::vector<std::string>& |
| 66 | +InferRequest::RequestedOutputNames() |
| 67 | +{ |
| 68 | + return requested_output_names_; |
| 69 | +} |
| 70 | + |
| 71 | +const std::string& |
| 72 | +InferRequest::ModelName() |
| 73 | +{ |
| 74 | + return model_name_; |
| 75 | +} |
| 76 | + |
| 77 | +int64_t |
| 78 | +InferRequest::ModelVersion() |
| 79 | +{ |
| 80 | + return model_version_; |
| 81 | +} |
| 82 | + |
| 83 | +void |
| 84 | +InferRequest::SaveToSharedMemory( |
| 85 | + std::unique_ptr<SharedMemory>& shm_pool, Request* request_shm) |
| 86 | +{ |
| 87 | + request_shm->correlation_id = this->CorrelationId(); |
| 88 | + off_t id_offset; |
| 89 | + SaveStringToSharedMemory(shm_pool, id_offset, this->RequestId().c_str()); |
| 90 | + request_shm->id = id_offset; |
| 91 | + request_shm->requested_output_count = this->RequestedOutputNames().size(); |
| 92 | + off_t requested_output_names_offset; |
| 93 | + off_t* requested_output_names; |
| 94 | + shm_pool->Map( |
| 95 | + (char**)&requested_output_names, |
| 96 | + sizeof(off_t) * request_shm->requested_output_count, |
| 97 | + requested_output_names_offset); |
| 98 | + |
| 99 | + request_shm->requested_output_names = requested_output_names_offset; |
| 100 | + size_t i = 0; |
| 101 | + for (auto& requested_output_name : requested_output_names_) { |
| 102 | + SaveStringToSharedMemory( |
| 103 | + shm_pool, requested_output_names[i], requested_output_name.c_str()); |
| 104 | + i++; |
| 105 | + } |
| 106 | + |
| 107 | + request_shm->requested_input_count = this->Inputs().size(); |
| 108 | + request_shm->model_version = this->model_version_; |
| 109 | + SaveStringToSharedMemory( |
| 110 | + shm_pool, request_shm->model_name, this->model_name_.c_str()); |
| 111 | +} |
| 112 | + |
| 113 | +std::unique_ptr<InferRequest> |
| 114 | +InferRequest::LoadFromSharedMemory( |
| 115 | + std::unique_ptr<SharedMemory>& shm_pool, off_t request_offset) |
| 116 | +{ |
| 117 | + Request* request; |
| 118 | + shm_pool->MapOffset((char**)&request, request_offset); |
| 119 | + |
| 120 | + char* id = nullptr; |
| 121 | + LoadStringFromSharedMemory(shm_pool, request->id, id); |
| 122 | + |
| 123 | + uint32_t requested_input_count = request->requested_input_count; |
| 124 | + |
| 125 | + std::vector<std::shared_ptr<PbTensor>> py_input_tensors; |
| 126 | + for (size_t input_idx = 0; input_idx < requested_input_count; ++input_idx) { |
| 127 | + std::shared_ptr<PbTensor> pb_input_tensor = PbTensor::LoadFromSharedMemory( |
| 128 | + shm_pool, request->inputs + sizeof(Tensor) * input_idx); |
| 129 | + py_input_tensors.emplace_back(std::move(pb_input_tensor)); |
| 130 | + } |
| 131 | + |
| 132 | + std::vector<std::string> requested_output_names; |
| 133 | + uint32_t requested_output_count = request->requested_output_count; |
| 134 | + off_t* output_names; |
| 135 | + shm_pool->MapOffset((char**)&output_names, request->requested_output_names); |
| 136 | + |
| 137 | + for (size_t output_idx = 0; output_idx < requested_output_count; |
| 138 | + ++output_idx) { |
| 139 | + char* output_name = nullptr; |
| 140 | + LoadStringFromSharedMemory(shm_pool, output_names[output_idx], output_name); |
| 141 | + requested_output_names.emplace_back(output_name); |
| 142 | + } |
| 143 | + |
| 144 | + char* model_name; |
| 145 | + LoadStringFromSharedMemory(shm_pool, request->model_name, model_name); |
| 146 | + return std::make_unique<InferRequest>( |
| 147 | + id, request->correlation_id, std::move(py_input_tensors), |
| 148 | + requested_output_names, model_name, request->model_version); |
| 149 | +} |
| 150 | + |
| 151 | +}}} // namespace triton::backend::python |
0 commit comments