graph_rt.cpp 25.5 KB
Newer Older
M
Megvii Engine Team 已提交
1 2 3 4
/**
 * \file imperative/python/src/graph_rt.cpp
 * MegEngine is Licensed under the Apache License, Version 2.0 (the "License")
 *
5
 * Copyright (c) 2014-2021 Megvii Inc. All rights reserved.
M
Megvii Engine Team 已提交
6 7 8 9 10 11
 *
 * 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
#include "./graph_rt.h"

14
#include "megbrain/graph/cg.h"
15
#include "megbrain/serialization/serializer.h"
16
#include "megbrain/imperative/opr_utility.h"
M
Megvii Engine Team 已提交
17
#include "megbrain/opr/io.h"
18
#include "megbrain/opr/utility.h"
19 20 21
#include "megbrain/opr/basic_arith.h"
#include "megbrain/imperative.h"
#include "./helper.h"
22
#include "megbrain/plugin/profiler.h"
23
#include "./common.h"
24 25
#include "megbrain/gopt/inference.h"

26 27 28 29 30

namespace py = pybind11;

using namespace mgb;
using namespace imperative;
31
namespace ser = mgb::serialization;
32

33 34
using _OptimizeForInferenceOptions = mgb::gopt::OptimizeForInferenceOptions;
using _LayoutTransform = _OptimizeForInferenceOptions::LayoutTransform;
35
using _AlgoStrategy = opr::mixin::AlgoChooserHelper::ExecutionPolicy::Strategy;
36

37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
namespace {
class _CompGraphProfilerImpl {
    std::shared_ptr<ComputingGraph> m_comp_graph;
    GraphProfiler m_profiler;
    public:
        _CompGraphProfilerImpl(std::shared_ptr<ComputingGraph> cg):
            m_comp_graph{cg},
            m_profiler{m_comp_graph.get()}
        {
        }

        std::string _get_result() {
            auto json = m_profiler.to_json_full(
                    m_comp_graph->current_comp_seq());
            return json->to_string();
        }
};
54 55 56 57 58 59 60

struct WeakRendezvousArray:
    public std::vector<std::weak_ptr<RendezvousBase>>,
    public UserDataContainer::UserData {
    MGB_TYPEINFO_OBJ_DECL;
};
MGB_TYPEINFO_OBJ_IMPL(WeakRendezvousArray);
61
}
62 63 64 65 66
#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)
67
        .def(py::init([](){return Rendezvous<T>::make();}))
68 69
        .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
70
        .def("drop", &Rendezvous<T>::drop)
71 72 73 74 75
        .def("reset", &Rendezvous<T>::reset)
        .def("set_exception", [](Rendezvous<T>& r, std::string&& message) {
            r.set_exception(std::make_exception_ptr(
                    std::runtime_error(std::move(message))));
        });
76 77 78
}

using TensorAttr = LogicalTensorDesc;
M
Megvii Engine Team 已提交
79
using HostNDWithEvent = std::pair<HostTensorND, std::shared_ptr<CompNode::Event>>;
80

81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
std::vector<mgb::cg::VarNode*>  _replace_vars(const std::vector<mgb::cg::VarNode*>& repl_src,
                                 const std::vector<mgb::cg::VarNode*>& repl_dst,
                                 const std::vector<mgb::cg::VarNode*>& vars) {
        mgb::ThinHashMap<SymbolVar, SymbolVar> varmap;
        for (size_t i = 0; i < repl_src.size(); ++i) {
            varmap[SymbolVar(repl_src[i])] = SymbolVar(repl_dst[i]);
        }
        SymbolVarArray symvars(vars.begin(), vars.end());
        auto sym_result = mgb::cg::replace_vars(symvars, varmap);
        std::vector<mgb::cg::VarNode*> result;
        for (auto symvar : sym_result){
            result.push_back(symvar.node());
        }
        return result;
    }

typedef std::vector<mgb::cg::OperatorNodeBase*> OperatorArray;
std::vector<mgb::cg::VarNode*> _replace_oprs(const OperatorArray& repl_src,
                                 const OperatorArray& repl_dst,
                                 const std::vector<mgb::cg::VarNode*>& vars) {
        mgb::ThinHashMap<mgb::cg::OperatorNodeBase*, mgb::cg::OperatorNodeBase*>
                oprmap;
        for (size_t i = 0; i < repl_src.size(); ++i) {
            oprmap[repl_src[i]] = repl_dst[i];
        }
        const SymbolVarArray symvars(vars.begin(), vars.end());
        auto sym_result = mgb::cg::replace_oprs(symvars, oprmap);
        std::vector<mgb::cg::VarNode*> result;
        for (auto symvar : sym_result){
            result.push_back(symvar.node());
        }
        return result;
    }



void _set_priority_to_id(const std::vector<mgb::cg::VarNode*>& dest_vars) {
        auto on_opr = [](mgb::cg::OperatorNodeBase* opr) {
            if (opr->node_prop().attribute().priority == 0) {
                opr->node_prop().attribute().priority = opr->id();
            }
        };
        mgb::cg::DepOprIter dep_iter{on_opr};
        for (const auto& var : dest_vars) {
            dep_iter.add(SymbolVar(var));
        }
}



131
void init_graph_rt(py::module m) {
132 133 134

   static const std::unique_ptr<mgb::OprFootprint> _imperative_sm_opr_footprint_ptr{std::make_unique<mgb::OprFootprint>()};

135 136
    def_rendezvous<DeviceTensorND>(m, "DeviceTensorNDRendezvous");

M
Megvii Engine Team 已提交
137 138
    def_rendezvous<HostNDWithEvent>(m, "HostTensorNDRendezvous");

139 140 141 142 143
    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();})
144 145
        .def_property("name", py::overload_cast<>(&VarNode::name, py::const_),
                      py::overload_cast<std::string>(&VarNode::name))
146
        .def_property_readonly("dtype", [](cg::VarNode* v) {return v->dtype();})
M
Megvii Engine Team 已提交
147 148 149 150
        .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();
                return mgr.infer_shape_fallible(v);
151 152 153 154 155 156 157 158 159 160 161 162 163
            })
        .def_property_readonly("value", [](cg::VarNode* v) -> py::object {
                auto&& mgr = v->owner_graph()->static_infer_manager();
                auto&& type = mgr.get_infer_type(v);
                using InferType = cg::static_infer::InferType;
                if (!(type.value & (InferType::CONST | InferType::RT_STATIC))) {
                    return py::none();
                }
                auto* val = mgr.infer_value_fallible(v);
                if (!val) {
                    return py::none();
                }
                return py::cast(*val).attr("numpy")();
164 165 166
            })
        .def_property_readonly("id",[](cg::VarNode* v){
            return (v->id());
167 168 169
        })
        .def("__repr__", [](cg::VarNode* v) {
            return "Var:" + v->name();
170
        });
171 172 173

    py::class_<cg::OperatorNodeBase, GraphNodePtr<cg::OperatorNodeBase>>(m, "OperatorNode")
        .def_property_readonly("graph", [](cg::OperatorNodeBase* opr) {return opr->owner_graph();})
174 175
        .def_property("name", py::overload_cast<>(&cg::OperatorNodeBase::name, py::const_),
                      py::overload_cast<std::string>(&cg::OperatorNodeBase::name))
176 177 178 179
        .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
180
                return to_tuple(opr->usable_output());
181 182 183 184 185 186 187 188 189
            })
        .def_property_readonly("id",[](cg::OperatorNodeBase* opr){
            return opr->id();
        })
        .def_property_readonly("params",[](cg::OperatorNodeBase* opr){
            return _imperative_sm_opr_footprint_ptr->calc_footprint(opr).param->to_string();
        })
        .def_property_readonly("type",[](cg::OperatorNodeBase* opr){
            return opr->dyn_typeinfo()->name;
190 191 192
        })
        .def("__repr__", [](cg::OperatorNodeBase* opr){
            return "Opr:" + opr->name();
193 194
        });

195 196
    py::class_<cg::AsyncExecutable>(m, "AsyncExecutable")
        .def("execute", &cg::AsyncExecutable::execute, py::call_guard<py::gil_scoped_release>())
197
        .def("wait", &cg::AsyncExecutable::wait, py::call_guard<py::gil_scoped_release>())
198
        .def("get_prev_exec_time", &cg::AsyncExecutable::get_prev_exec_time, py::call_guard<py::gil_scoped_release>())
199 200 201 202 203 204 205 206 207 208 209 210 211 212
        // only used for exception handle
        .def_property_readonly("_all_rendezvous", [](cg::AsyncExecutable* exec) {
            auto ud = exec->owner_graph()->options().user_data
                        .get_user_data<WeakRendezvousArray>();
            std::vector<std::shared_ptr<RendezvousBase>> ret;
            if (ud.second) {
                for (auto&& r: *ud.first[0]) {
                    if (auto p = r.lock()) {
                        ret.emplace_back(std::move(p));
                    }
                }
            }
            return ret;
        });
213 214 215 216 217 218 219 220 221 222 223 224 225

    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));

226 227 228 229 230 231
    py::class_<_CompGraphProfilerImpl, std::shared_ptr<_CompGraphProfilerImpl>>(m, "GraphProfiler")
        .def(py::init([](std::shared_ptr<ComputingGraph> graph) {
                return std::make_shared<_CompGraphProfilerImpl>(graph);
                }))
        .def("get", [](_CompGraphProfilerImpl& profiler) { return profiler._get_result(); });

232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261
    auto GraphOptimizeOptions = py::class_<_OptimizeForInferenceOptions>(m, "GraphOptimizeOptions")
        .def(py::init())
        .def_readwrite("f16_io_f32_comp", &_OptimizeForInferenceOptions::f16_io_f32_comp)
        .def_readwrite("f16_io_comp", &_OptimizeForInferenceOptions::f16_io_comp)
        .def_readwrite("fuse_conv_bias_nonlinearity", &_OptimizeForInferenceOptions::fuse_conv_bias_nonlinearity)
        .def_readwrite("fuse_conv_bias_with_z", &_OptimizeForInferenceOptions::fuse_conv_bias_with_z)
        .def_readwrite("layout_transform", &_OptimizeForInferenceOptions::layout_transform)
        ;

    py::enum_<_LayoutTransform>(GraphOptimizeOptions, "LayoutTransform")
        .value("DEFAULT", _LayoutTransform::DEFAULT)
        .value("NCHW4", _LayoutTransform::NCHW4)
        .value("NHWCD4", _LayoutTransform::NHWCD4)
        .value("NCHW88", _LayoutTransform::NCHW88)
        .value("NCHW44", _LayoutTransform::NCHW44)
        .value("NCHW44_DOT", _LayoutTransform::NCHW44_DOT)
        .value("NCHW32", _LayoutTransform::NCHW32)
        .value("CHWN4", _LayoutTransform::CHWN4)
        .export_values()
        ;

    m.def("optimize_for_inference", [](const VarNodeArray& dest_vars, const _OptimizeForInferenceOptions& opt) {
        SymbolVarArray symvars(dest_vars.begin(), dest_vars.end());
        auto res_symvars = mgb::gopt::optimize_for_inference(symvars, opt);
        VarNodeArray vars;
        for (auto& si: res_symvars)
            vars.push_back(si.node());
        return vars;
    });

262 263
    m.def("modify_opr_algo_strategy_inplace", [](const VarNodeArray& dest_vars,
                                                 const std::string& strategy) {
264
        _AlgoStrategy stg;
265 266 267 268 269
        const std::unordered_map<std::string, std::function<void()>> m{
                {"HEURISTIC", [&]() { stg = _AlgoStrategy::HEURISTIC; }},
                {"PROFILE", [&]() { stg = _AlgoStrategy::PROFILE; }},
                {"REPRODUCIBLE", [&]() { stg = _AlgoStrategy::REPRODUCIBLE; }},
                {"OPTMIZED", [&]() { stg = _AlgoStrategy::OPTMIZED; }},
270 271 272 273 274 275 276
        };
        auto it = m.find(strategy);
        mgb_assert(it != m.end(), "Invalid strategy string!");
        it->second();
        mgb::gopt::modify_opr_algo_strategy_inplace(dest_vars, stg);
    });

277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307
    m.def("get_info_for_strip", [](const std::vector<VarNode*>& dest_vars) {
        std::unordered_set<const char*> opr_types, dtype_names, elemwise_modes;
        auto on_opr = [&](cg::OperatorNodeBase *opr) {
            if (ser::GraphDumper::should_remove_in_dump(opr))
                return;
            opr_types.insert(opr->dyn_typeinfo()->name);
            for (auto i : opr->output())
                dtype_names.insert(i->dtype().name());
            if (opr->same_type<opr::Elemwise>()) {
                auto mode = opr->cast_final<opr::Elemwise>().param().mode;
                elemwise_modes.insert(
                        megdnn::Elemwise::ModeTrait::from_mode(mode).name);
            }
        };
        cg::DepOprIter opr_iter{on_opr};
        for (auto i : dest_vars)
            opr_iter.add(i->owner_opr());

        auto to_json = [](const std::unordered_set<const char*> &v) {
            std::vector<std::string> vs(v.begin(), v.end());
            std::sort(vs.begin(), vs.end());
            auto ret = json::Array::make();
            for (auto &&i : vs)
                ret->add(json::String::make(i));
            return ret;
        };

        return json::Object::make({
            {"opr_types", to_json(opr_types)},
            {"dtypes", to_json(dtype_names)},
            {"elemwise_modes", to_json(elemwise_modes)},
308
        })->to_string();
309 310 311 312 313
    });

    m.def("dump_graph", [](
        const std::vector<VarNode*>& dest_vars,
        int keep_var_name,
314
        bool keep_opr_name,
315 316 317 318 319 320 321 322 323 324 325 326
        bool keep_param_name,
        bool keep_opr_priority,
        py::list& stat,
        py::list& inputs,
        py::list& outputs,
        py::list& params
    ) {
        std::vector<uint8_t> buf;
        auto dumper = ser::GraphDumper::make(ser::OutputFile::make_vector_proxy(&buf));
        SymbolVarArray symvars(dest_vars.begin(), dest_vars.end());

        ser::GraphDumper::DumpConfig config{keep_var_name, keep_param_name,
327
                                       keep_opr_priority, keep_opr_name};
328 329 330 331 332 333 334 335 336 337 338

        auto rst = dumper->dump(symvars, config);
        for (auto i : rst.inputs) {
            inputs.append(py::cast(i));
        }
        for (auto i : rst.outputs) {
            outputs.append(py::cast(i));
        }
        for (auto i : rst.params) {
            params.append(py::cast(i));
        }
339 340 341
        auto rst_stat =
                std::vector{rst.nr_opr, rst.tot_bytes, rst.tensor_value_bytes,
                            static_cast<size_t>(rst.content_hash)};
342 343 344 345 346
        for (auto i : rst_stat) {
            stat.append(py::cast(i));
        }
        return py::bytes(reinterpret_cast<const char*>(&buf[0]), buf.size());
    });
347

348 349 350 351 352 353 354 355 356
    m.def("load_graph", [](
        std::string& buf,
        py::list& output_var_map,
        py::list& output_var_list
    ) {
        auto file = ser::InputFile::make_mem_proxy(buf.c_str(), buf.length());
        auto format = ser::GraphLoader::identify_graph_dump_format(*file);
        auto loader = ser::GraphLoader::make(std::move(file), format.val());
        ser::GraphLoader::LoadConfig config;
357
        auto rst = loader->load(config);
358 359
        for (auto i : rst.output_var_map) {
            output_var_map.append(py::make_tuple(i.first, i.second.node()));
360
        }
361 362
        for (auto i : rst.output_var_list) {
            output_var_list.append(i.node());
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377
        }
        std::unordered_map<HostTensorND*, const std::string*> tensor2name;
        for (const auto& pair : rst.tensor_map) {
            tensor2name[pair.second.get()] = &pair.first;
        }
        auto cb = [&tensor2name, graph=rst.graph](cg::OperatorNodeBase* opr) {
            if (!opr->same_type<opr::Host2DeviceCopy>())
                return;
            auto& h2d = opr->cast_final_safe<opr::Host2DeviceCopy>();
            auto it = tensor2name.find(h2d.host_data().get());
            mgb_throw_if(it == tensor2name.end(), GraphError,
                        "unbound Host2DeviceCopy in loaded graph");
            h2d.output(0)->name(*it->second);
        };
        cg::DepOprIter iter{cb};
378 379
        for (const auto& var : rst.output_var_list) {
            iter.add(var);
380 381 382 383 384
        }
        return rst.graph;

    });

385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403
#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)
404
        DEF_READWRITE(no_force_inplace)
405
        DEF_READWRITE(sublinear_mem_config)
406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
        // 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

429 430 431 432 433 434 435 436 437 438
#define CURRENT_CLASS cg::ComputingGraph::Options::SublinearMemConfig

    py::class_<cg::ComputingGraph::Options::SublinearMemConfig>(PyComputingGraphOptions, "SublinearMemConfig")
        DEF_READWRITE(thresh_nr_try)
        DEF_READWRITE(genetic_nr_iter)
        DEF_READWRITE(genetic_pool_size)
        DEF_READWRITE(lb_memory)
        DEF_READWRITE(num_worker);

#undef CURRENT_CLASS
439 440 441 442
    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());
443
            return to_tuple(OpDef::apply_on_var_node(def, vinputs));
444 445 446 447 448 449
        },
        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
450
                             const TensorShape& shape,
451
                             const std::vector<cg::VarNode*>& inputs,
452 453
                             cg::ComputingGraph* graph,
                             bool use_static_shape) {
454 455 456 457 458 459 460 461
        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);
462 463 464
        auto soutputs = opr::InputCallback::make(*graph, std::move(callback),
                                                 comp_node, dtype, shape,
                                                 sinputs, use_static_shape);
465 466 467 468 469 470 471 472
        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
473 474 475 476
    m.def("make_shared", [](cg::ComputingGraph* graph, const DeviceTensorND& data) {
            return opr::SharedDeviceTensor::make(*graph, std::make_shared<DeviceTensorND>(data)).node();
        });

477
    m.def("make_const", [](cg::ComputingGraph* graph, py::array data, CompNode cn, DType dtype, std::optional<std::string> name) {
M
Megvii Engine Team 已提交
478
            if (!cn.valid()) {
479
                cn = CompNode::load(get_default_device());
M
Megvii Engine Team 已提交
480
            }
481 482 483 484
            OperatorNodeConfig config(cn);
            if (name) {
                config.name(*name);
            }
M
Megvii Engine Team 已提交
485
            auto hv = npy::np2tensor(data.ptr(), npy::Meth::borrow(cn), dtype);
486 487
            return opr::ImmutableTensor::make(*graph, hv, config).node();
        }, py::arg(), py::arg(), py::arg(), py::arg(), py::arg() = py::none());
M
Megvii Engine Team 已提交
488

489
    m.def("make_h2d", [](cg::ComputingGraph& graph, CompNode cn, DType dtype, TensorShape shape, std::optional<std::string> name) {
490 491 492 493 494 495 496 497 498 499
            if (!cn.valid()) {
                throw py::type_error("device must be valid");
            }
            if (!dtype.valid()) {
                throw py::type_error("dtype must be valid");
            }
            OperatorNodeConfig config;
            if (name) {
                config.name(*name);
            }
500 501
            return opr::Host2DeviceCopy::make(graph, std::make_shared<HostTensorND>(cn, shape, dtype), config).node();
        }, py::arg(), py::arg(), py::arg(), py::arg() = py::none(), py::arg() = py::none());
502

503 504 505 506
    m.def("_replace_vars", &_replace_vars,py::arg(),py::arg(),py::arg());
    m.def("_replace_oprs", &_replace_oprs,py::arg(),py::arg(),py::arg());
    m.def("_set_priority_to_id",&_set_priority_to_id,py::arg());

507 508 509
    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
510
                                             const TensorShape& shape,
511
                                             const std::vector<cg::VarNode*>& inputs,
512 513 514 515 516
                                             cg::ComputingGraph* graph,
                                             bool use_static_shape) {
            return input_callback(
                [f=std::move(callback)](){py::gil_scoped_acquire _; return f();},
                comp_node, dtype, shape, inputs, graph, use_static_shape);
517
        },
518 519
        py::arg(), py::arg(), py::arg(), py::arg() = py::none(), py::arg() = py::tuple(),
        py::arg("graph") = py::none(), py::arg("use_static_shape") = false);
520 521 522 523

    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
524
                                             const TensorShape& shape,
525
                                             const std::vector<cg::VarNode*>& inputs,
526 527
                                             cg::ComputingGraph* graph,
                                             bool use_static_shape) {
528 529 530
            auto f = [p]() -> DeviceTensorND {
                return p->get();
            };
531
            return input_callback(std::move(f), comp_node, dtype, shape, inputs, graph, use_static_shape);
532
        },
533 534
        py::arg(), py::arg(), py::arg(), py::arg() = py::none(), py::arg() = py::tuple(), 
        py::arg("graph") = py::none(), py::arg("use_static_shape") = false);
535

536 537 538 539 540 541 542 543
    auto output_callback = [](auto callback, const std::vector<cg::VarNode*>& inputs,
            std::shared_ptr<RendezvousBase> r = {}, bool borrow = false, bool prefer_host_value = false) {
        if (r) {
            mgb_assert(inputs.size());
            auto cg = inputs[0]->owner_graph();
            cg->options().user_data.get_user_data_or_create<WeakRendezvousArray>()
                    ->emplace_back(r);
        }
544 545 546 547 548
        SymbolVarArray sinputs;
        for (auto i : inputs) {
            sinputs.emplace_back(i);
        }
        static_assert(!std::is_reference<decltype(callback)>::value);
549
        opr::OutputCallback::Param param{std::move(callback), borrow, prefer_host_value};
550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567
        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));
        };
568
        return output_callback(std::move(f), std::move(inputs), p);
569 570
    });

M
Megvii Engine Team 已提交
571 572 573 574 575 576 577 578
    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));
        };
579
        return output_callback(std::move(f), std::move(inputs), p, true, true);
M
Megvii Engine Team 已提交
580 581
    });

582 583 584 585
    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()});
        };
586
        return output_callback(std::move(f), std::move(inputs), p, true);
587
    });
588 589 590 591 592 593 594 595 596 597 598 599

    m.def("virtual_dep", [](std::vector<cg::VarNode*> inputs, std::string device) {
        auto&& graph = inputs[0]->owner_graph();
        VarNodeArray inps(inputs.begin(), inputs.end());
        cg::OperatorNodeConfig config;
        if (device.length() > 0) {
            config.comp_node(CompNode::load(device));
        }
        cg::OperatorNodeBase* opr = graph->insert_opr(
                std::make_unique<mgb::opr::VirtualDep>(inps, config));
        return opr;
    });
600
}