graph_rt.cpp 10.9 KB
Newer Older
M
Megvii Engine Team 已提交
1 2 3 4 5 6 7 8 9 10 11
/**
 * \file imperative/python/src/graph_rt.cpp
 * MegEngine is Licensed under the Apache License, Version 2.0 (the "License")
 *
 * Copyright (c) 2014-2020 Megvii Inc. All rights reserved.
 *
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT ARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 */

12 13 14
#include "./graph_rt.h"

#include "megbrain/imperative/opr_utility.h"
M
Megvii Engine Team 已提交
15
#include "megbrain/opr/io.h"
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
#include "megbrain/opr/basic_arith.h"
#include "megbrain/imperative.h"
#include "./helper.h"

namespace py = pybind11;

using namespace mgb;
using namespace imperative;

#define DEF_READWRITE(name) .def_readwrite(#name, &CURRENT_CLASS::name)

template<typename T>
auto def_rendezvous(py::object m, const char* name) {
    return py::class_<Rendezvous<T>, std::shared_ptr<Rendezvous<T>>>(m, name)
        .def(py::init([](){return std::make_shared<Rendezvous<T>>();}))
        .def("set", [](Rendezvous<T>& r, T v) {r.set(std::move(v));})
        .def("get", [](Rendezvous<T>& r) {return r.get();}, py::call_guard<py::gil_scoped_release>())
M
Megvii Engine Team 已提交
33
        .def("drop", &Rendezvous<T>::drop)
34 35 36 37
        .def("reset", &Rendezvous<T>::reset);
}

using TensorAttr = LogicalTensorDesc;
M
Megvii Engine Team 已提交
38
using HostNDWithEvent = std::pair<HostTensorND, std::shared_ptr<CompNode::Event>>;
39 40 41 42

void init_graph_rt(py::module m) {
    def_rendezvous<DeviceTensorND>(m, "DeviceTensorNDRendezvous");

M
Megvii Engine Team 已提交
43 44
    def_rendezvous<HostNDWithEvent>(m, "HostTensorNDRendezvous");

45 46 47 48 49
    def_rendezvous<TensorAttr>(m, "TensorAttrRendezvous");

    py::class_<cg::VarNode, GraphNodePtr<cg::VarNode>>(m, "VarNode")
        .def_property_readonly("owner", [](cg::VarNode* v) {return v->owner_opr();})
        .def_property_readonly("graph", [](cg::VarNode* v) {return v->owner_graph();})
M
Megvii Engine Team 已提交
50
        .def_property_readonly("name", py::overload_cast<>(&VarNode::name, py::const_))
51
        .def_property_readonly("dtype", [](cg::VarNode* v) {return v->dtype();})
M
Megvii Engine Team 已提交
52 53 54 55 56 57 58 59 60 61
        .def_property_readonly("comp_node", [](cg::VarNode* v) {return v->comp_node();})
        .def_property_readonly("shape", [](cg::VarNode* v) -> const TensorShape* {
                auto&& mgr = v->owner_graph()->static_infer_manager();
                auto&& type = mgr.get_infer_type(v);
                using InferType = cg::static_infer::InferType;
                if (!(type.shape & (InferType::CONST | InferType::RT_STATIC))) {
                    return nullptr;
                }
                return mgr.infer_shape_fallible(v);
            });
62 63 64

    py::class_<cg::OperatorNodeBase, GraphNodePtr<cg::OperatorNodeBase>>(m, "OperatorNode")
        .def_property_readonly("graph", [](cg::OperatorNodeBase* opr) {return opr->owner_graph();})
M
Megvii Engine Team 已提交
65
        .def_property_readonly("name", py::overload_cast<>(&cg::OperatorNodeBase::name, py::const_))
66 67 68 69
        .def_property_readonly("inputs", [](cg::OperatorNodeBase* opr) {
                return to_tuple(opr->input());
            })
        .def_property_readonly("outputs", [](cg::OperatorNodeBase* opr) {
M
Megvii Engine Team 已提交
70
                return to_tuple(opr->usable_output());
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
            });

    py::class_<cg::AsyncExecutable>(m, "AsyncExecutable")
        .def("execute", &cg::AsyncExecutable::execute, py::call_guard<py::gil_scoped_release>())
        .def("wait", &cg::AsyncExecutable::wait, py::call_guard<py::gil_scoped_release>());

    auto PyComputingGraph = py::class_<cg::ComputingGraph, std::shared_ptr<cg::ComputingGraph>>(m, "ComputingGraph")
        .def(py::init(py::overload_cast<>(&cg::ComputingGraph::make)))
        .def("compile", [](cg::ComputingGraph& graph, const std::vector<cg::VarNode*>& dest_vars) {
                mgb_assert(!dest_vars.empty());
                cg::ComputingGraph::OutputSpec spec;
                for (auto v : dest_vars) {
                    spec.emplace_back(v, nullptr);
                }
                return graph.compile(spec);
            })
        .def_property_readonly("options", py::overload_cast<>(&cg::ComputingGraph::options));

#define CURRENT_CLASS cg::ComputingGraph::Options

    auto PyComputingGraphOptions = py::class_<cg::ComputingGraph::Options>(PyComputingGraph, "Options")
        // DEF_READWRITE(opr_attribute)
        DEF_READWRITE(seq_opt)
        DEF_READWRITE(graph_opt)
        DEF_READWRITE(graph_opt_level)
        DEF_READWRITE(log_level)
        DEF_READWRITE(async_exec_level)
        DEF_READWRITE(force_dynamic_alloc)
        DEF_READWRITE(var_sanity_check_first_run)
        DEF_READWRITE(allocate_static_mem_after_graph_compile)
        DEF_READWRITE(fake_next_exec)
        DEF_READWRITE(enable_sublinear_memory_opt)
        DEF_READWRITE(no_profiling_on_shape_change)
        DEF_READWRITE(enable_var_mem_defragment)
        DEF_READWRITE(enable_grad_var_static_reshape)
        DEF_READWRITE(enable_memory_swap)
        DEF_READWRITE(comp_node_seq_record_level)
108
        DEF_READWRITE(no_force_inplace)
109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
        // DEF_READWRITE(eager_evaluation)
        // DEF_READWRITE(imperative_proxy_graph)
        // DEF_READWRITE(extra_vardeps)
        // DEF_READWRITE(user_data)
        ;

#undef CURRENT_CLASS
#define CURRENT_CLASS cg::ComputingGraph::Options::SeqOpt

    py::class_<cg::ComputingGraph::Options::SeqOpt>(PyComputingGraphOptions, "SeqOpt")
        DEF_READWRITE(enable_mem_plan_opt)
        DEF_READWRITE(enable_mem_reuse_alloc)
        DEF_READWRITE(enable_seq_comp_node_opt);

#undef CURRENT_CLASS
#define CURRENT_CLASS cg::ComputingGraph::Options::GraphOpt

    py::class_<cg::ComputingGraph::Options::GraphOpt>(PyComputingGraphOptions, "GraphOpt")
        DEF_READWRITE(jit)
        DEF_READWRITE(tensorrt);

#undef CURRENT_CLASS

    auto common = rel_import("common", m, 1);

    common.def("invoke_op", [](const OpDef& def, const std::vector<cg::VarNode*> inputs, cg::ComputingGraph* graph) {
            cg::VarNodeArray vinputs(inputs.begin(), inputs.end());
            auto opr = OpDef::apply_on_var_node(def, vinputs);
M
Megvii Engine Team 已提交
137
            auto outputs = opr->usable_output();
138 139 140 141 142 143 144
            return to_tuple(outputs);
        },
        py::arg(), py::arg(), py::arg("graph") = py::none());

    auto input_callback = [](auto callback,
                             const CompNode& comp_node,
                             const DType& dtype,
M
Megvii Engine Team 已提交
145
                             const TensorShape& shape,
146 147 148 149 150 151 152 153 154 155
                             const std::vector<cg::VarNode*>& inputs,
                             cg::ComputingGraph* graph) {
        if (!graph) {
            graph = inputs[0]->owner_graph();
        }
        SymbolVarArray sinputs;
        for (auto i : inputs) {
            sinputs.emplace_back(i);
        }
        static_assert(!std::is_reference<decltype(callback)>::value);
M
Megvii Engine Team 已提交
156
        auto soutputs = opr::InputCallback::make(*graph, std::move(callback), comp_node, dtype, shape, sinputs);
157 158 159 160 161 162 163 164
        std::vector<VarNode*> outputs;
        outputs.reserve(soutputs.size());
        for (auto i : soutputs) {
            outputs.push_back(i.node());
        }
        return outputs;
    };

M
Megvii Engine Team 已提交
165 166 167 168 169 170 171 172 173 174 175 176
    m.def("make_shared", [](cg::ComputingGraph* graph, const DeviceTensorND& data) {
            return opr::SharedDeviceTensor::make(*graph, std::make_shared<DeviceTensorND>(data)).node();
        });

    m.def("make_const", [](cg::ComputingGraph* graph, py::array data, CompNode cn, DType dtype) {
            if (!cn.valid()) {
                throw py::type_error("device must not be None");
            }
            auto hv = npy::np2tensor(data.ptr(), npy::Meth::borrow(cn), dtype);
            opr::ImmutableTensor::make(*graph, hv, OperatorNodeConfig(cn)).node();
        });

177 178 179
    m.def("input_callback", [input_callback](std::function<DeviceTensorND(void)> callback,
                                             const CompNode& comp_node,
                                             const DType& dtype,
M
Megvii Engine Team 已提交
180
                                             const TensorShape& shape,
181 182
                                             const std::vector<cg::VarNode*>& inputs,
                                             cg::ComputingGraph* graph) {
M
Megvii Engine Team 已提交
183
            return input_callback([f=std::move(callback)](){py::gil_scoped_acquire _; return f();}, comp_node, dtype, shape, inputs, graph);
184
        },
M
Megvii Engine Team 已提交
185
        py::arg(), py::arg(), py::arg(), py::arg() = py::none(), py::arg() = py::tuple(), py::arg("graph") = py::none());
186 187 188 189

    m.def("input_callback", [input_callback](std::shared_ptr<Rendezvous<DeviceTensorND>> p,
                                             const CompNode& comp_node,
                                             const DType& dtype,
M
Megvii Engine Team 已提交
190
                                             const TensorShape& shape,
191 192 193 194 195
                                             const std::vector<cg::VarNode*>& inputs,
                                             cg::ComputingGraph* graph) {
            auto f = [p]() -> DeviceTensorND {
                return p->get();
            };
M
Megvii Engine Team 已提交
196
            return input_callback(std::move(f), comp_node, dtype, shape, inputs, graph);
197
        },
M
Megvii Engine Team 已提交
198
        py::arg(), py::arg(), py::arg(), py::arg() = py::none(), py::arg() = py::tuple(), py::arg("graph") = py::none());
199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227

    auto output_callback = [](auto callback, const std::vector<cg::VarNode*>& inputs, bool borrow = false) {
        SymbolVarArray sinputs;
        for (auto i : inputs) {
            sinputs.emplace_back(i);
        }
        static_assert(!std::is_reference<decltype(callback)>::value);
        opr::OutputCallback::Param param{std::move(callback), borrow};
        auto output = opr::OutputCallback::make(std::move(param), sinputs);
        return output.node();
    };

    m.def("output_callback", [output_callback](std::function<void(DeviceTensorND)> callback, std::vector<cg::VarNode*> inputs) {
        auto f = [f=std::move(callback)](DeviceTensorND dv) {
            auto task = [f=std::move(f), dv=std::move(dv)]() {
                f(dv);
            };
            py_task_q.add_task(std::move(task));
        };
        return output_callback(std::move(f), std::move(inputs));
    });

    m.def("output_callback", [output_callback](std::shared_ptr<Rendezvous<DeviceTensorND>> p, std::vector<cg::VarNode*> inputs) {
        auto f = [p](DeviceTensorND dv) {
            p->set(std::move(dv));
        };
        return output_callback(std::move(f), std::move(inputs));
    });

M
Megvii Engine Team 已提交
228 229 230 231 232 233 234 235 236 237 238
    m.def("value_output_callback", [output_callback](std::shared_ptr<Rendezvous<HostNDWithEvent>> p, std::vector<cg::VarNode*> inputs) {
        auto f = [p](DeviceTensorND dv) {
            HostNDWithEvent hv_with_event;
            hv_with_event.first.copy_from(dv);
            hv_with_event.second = dv.comp_node().create_event();
            hv_with_event.second->record();
            p->set(std::move(hv_with_event));
        };
        return output_callback(std::move(f), std::move(inputs), true);
    });

239 240 241 242 243 244 245
    m.def("attr_output_callback", [output_callback](std::shared_ptr<Rendezvous<TensorAttr>> p, std::vector<cg::VarNode*> inputs) {
        auto f = [p](DeviceTensorND dv) {
            p->set(TensorAttr{TensorLayout{dv.shape(), dv.dtype()}, dv.comp_node()});
        };
        return output_callback(std::move(f), std::move(inputs), true);
    });
}