Nigel
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

!47778 Clean code warning. 

Merge pull request !47778 from Margaret_wangrui/clean_code
This commit is contained in:
i-robot 2023-01-14 08:30:44 +00:00 committed by Gitee
parent e7a35cf5e2 8b65d3092f
commit c32a8cd837
No known key found for this signature in database
GPG Key ID: 173E9B9CA92EEF8F
9 changed files with 39 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
mindspore/ccsrc/backend/common/pass/convert_tuple_input_to_dynamic_input.cc
View File

@ -1,5 +1,5 @@
/** /**
* Copyright 2020-2022 Huawei Technologies Co., Ltd * Copyright 2020-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -45,7 +45,7 @@ int64_t SplitTupleInputs(const FuncGraphPtr &graph, const AnfNodePtr &tuple_inpu
MS_EXCEPTION_IF_NULL(dyn_input_node); MS_EXCEPTION_IF_NULL(dyn_input_node);
// Handle tuple nested scenes. // Handle tuple nested scenes.
if (dyn_input_node->isa<CNode>() && common::AnfAlgo::CheckPrimitiveType(dyn_input_node, prim::kPrimMakeTuple)) { if (dyn_input_node->isa<CNode>() && common::AnfAlgo::CheckPrimitiveType(dyn_input_node, prim::kPrimMakeTuple)) {
input_size += SplitTupleInputs(graph, dyn_input_node, plant_inputs); input_size += LongToSize(SplitTupleInputs(graph, dyn_input_node, plant_inputs));
continue; continue;
} }
(void)plant_inputs->emplace_back(dyn_input_node); (void)plant_inputs->emplace_back(dyn_input_node);

10
mindspore/ccsrc/frontend/operator/composite/composite.cc
View File

@ -1,7 +1,7 @@
/** /**
* This is the C++ adaptation and derivative work of Myia (https://github.com/mila-iqia/myia/). * This is the C++ adaptation and derivative work of Myia (https://github.com/mila-iqia/myia/).
* *
* Copyright 2019-2022 Huawei Technologies Co., Ltd * Copyright 2019-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -145,7 +145,7 @@ AnfNodePtr HyperMap::FullMake(const std::shared_ptr<List> &type, const FuncGraph
MS_LOG(EXCEPTION) << "The lists in HyperMap should have the same length. " << oss.str(); MS_LOG(EXCEPTION) << "The lists in HyperMap should have the same length. " << oss.str();
} }
// cannot use shared_from_base() also known as this, as it will make a reference cycle on // Cannot use shared_from_base() also known as this, as it will make a reference cycle on
// hypermap and graph generated, it will cause memory leak. // hypermap and graph generated, it will cause memory leak.
auto fn_rec = NewValueNode(std::make_shared<HyperMap>(*this)); auto fn_rec = NewValueNode(std::make_shared<HyperMap>(*this));
constexpr size_t kPrimHoldLen = 1; constexpr size_t kPrimHoldLen = 1;
@ -205,7 +205,7 @@ AnfNodePtr HyperMap::FullMake(const std::shared_ptr<Tuple> &type, const FuncGrap
MS_LOG(EXCEPTION) << "The length of tuples in HyperMap must be the same. " << oss.str(); MS_LOG(EXCEPTION) << "The length of tuples in HyperMap must be the same. " << oss.str();
} }
// cannot use shared_from_base() also known as this, as it will make a reference cycle on // Cannot use shared_from_base() also known as this, as it will make a reference cycle on
// hypermap and graph generated, it will cause memory leak. // hypermap and graph generated, it will cause memory leak.
auto fn_rec = NewValueNode(std::make_shared<HyperMap>(*this)); auto fn_rec = NewValueNode(std::make_shared<HyperMap>(*this));
constexpr size_t kPrimHoldLen = 1; constexpr size_t kPrimHoldLen = 1;
@ -462,7 +462,7 @@ FuncGraphPtr PyExecuteGradient::GenerateFuncGraph(const AbstractBasePtrList &arg
// Make fprop first result, PyExecute's forward result. // Make fprop first result, PyExecute's forward result.
AnfNodePtr out = fg->NewCNodeInOrder(params); AnfNodePtr out = fg->NewCNodeInOrder(params);
// make fprop second result, PyExecute's backward function. // Make fprop second result, PyExecute's backward function.
FuncGraphPtr bprop = std::make_shared<FuncGraph>(); FuncGraphPtr bprop = std::make_shared<FuncGraph>();
ss.str(std::string()); ss.str(std::string());
@ -1252,7 +1252,7 @@ FuncGraphPtr VmapOperation::GenerateFuncGraph(const AbstractBasePtrList &args_sp
vmap_fg->set_flag(FUNC_GRAPH_FLAG_CORE, true); vmap_fg->set_flag(FUNC_GRAPH_FLAG_CORE, true);
vmap_fg->debug_info()->set_name(ss.str()); vmap_fg->debug_info()->set_name(ss.str());
// add parameter for `fn`, `in_axes` and `out_axes` respectively. // Add parameter for `fn`, `in_axes` and `out_axes` respectively.
ParameterPtr param_graph = vmap_fg->add_parameter(); ParameterPtr param_graph = vmap_fg->add_parameter();
(void)vmap_fg->add_parameter(); (void)vmap_fg->add_parameter();
(void)vmap_fg->add_parameter(); (void)vmap_fg->add_parameter();

4
mindspore/ccsrc/frontend/optimizer/ad/dfunctor.cc
View File

@ -1,5 +1,5 @@
/** /**
* Copyright 2020-2022 Huawei Technologies Co., Ltd * Copyright 2020-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -62,7 +62,7 @@ DFunctor::DFunctor(const FuncGraphPtr &primal_graph, const pipeline::ResourceBas
const auto &info = primal_graph->GetEffectInfo(); const auto &info = primal_graph->GetEffectInfo();
if (is_top_ && info.back_mem) { if (is_top_ && info.back_mem) {
// Add Umonad arg for top graph. // Add Umonad arg for top graph.
tape_->add_parameter(); (void)tape_->add_parameter();
} }
} }

6
mindspore/ccsrc/frontend/optimizer/clean.cc
View File

@ -1,7 +1,7 @@
/** /**
* This is the C++ adaptation and derivative work of Myia (https://github.com/mila-iqia/myia/). * This is the C++ adaptation and derivative work of Myia (https://github.com/mila-iqia/myia/).
* *
* Copyright 2019-2022 Huawei Technologies Co., Ltd * Copyright 2019-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -226,7 +226,7 @@ class SimplifyDataStructuresRewriter : public BaseRewriter {
} }
// DictGetItem --> PyExecute() // DictGetItem --> PyExecute()
AnfNodePtr RebuidDictGetItem(const CNodePtr &node) const { AnfNodePtr RebuildDictGetItem(const CNodePtr &node) const {
MS_EXCEPTION_IF_NULL(node); MS_EXCEPTION_IF_NULL(node);
// Inputs should be [dict_setitem, dict, item] // Inputs should be [dict_setitem, dict, item]
const size_t expect_inputs_size = 3; const size_t expect_inputs_size = 3;
@ -291,7 +291,7 @@ class SimplifyDataStructuresRewriter : public BaseRewriter {
AnfNodePtr ConvertDictGetItem(const CNodePtr &node) { AnfNodePtr ConvertDictGetItem(const CNodePtr &node) {
static const auto support_fallback_runtime = (common::GetEnv("MS_DEV_ENABLE_FALLBACK_RUNTIME") != "0"); static const auto support_fallback_runtime = (common::GetEnv("MS_DEV_ENABLE_FALLBACK_RUNTIME") != "0");
if (support_fallback_runtime && is_dict_output_) { if (support_fallback_runtime && is_dict_output_) {
return RebuidDictGetItem(node); return RebuildDictGetItem(node);
} }
return ConvertDictGetItemToTupleGetItem(node); return ConvertDictGetItemToTupleGetItem(node);
} }

27
mindspore/ccsrc/frontend/optimizer/irpass/cast_eliminate.cc
View File

@ -86,7 +86,7 @@ void CastSameTypeEliminater::Visit(const AnfNodePtr &node) {
} }
} }
bool TwoCastEliminater::CheckTwoTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2) { bool TwoCastEliminater::CheckTwoTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2) const {
auto type1_iter = type_map.find(type1); auto type1_iter = type_map.find(type1);
auto type2_iter = type_map.find(type2); auto type2_iter = type_map.find(type2);
if (type1_iter != type_map.end() && type2_iter != type_map.end()) { if (type1_iter != type_map.end() && type2_iter != type_map.end()) {
@ -96,7 +96,7 @@ bool TwoCastEliminater::CheckTwoTypes(const std::map<TypeId, int> &type_map, Typ
} }
bool TwoCastEliminater::CheckThreeTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2, bool TwoCastEliminater::CheckThreeTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2,
TypeId type3) { TypeId type3) const {
auto type1_iter = type_map.find(type1); auto type1_iter = type_map.find(type1);
auto type2_iter = type_map.find(type2); auto type2_iter = type_map.find(type2);
auto type3_iter = type_map.find(type3); auto type3_iter = type_map.find(type3);
@ -107,8 +107,8 @@ bool TwoCastEliminater::CheckThreeTypes(const std::map<TypeId, int> &type_map, T
} }
// {prim::kPrimCast, {prim::kPrimCast, X, Y}, T} -> {prim::kPrimCast, X, T} // {prim::kPrimCast, {prim::kPrimCast, X, Y}, T} -> {prim::kPrimCast, X, T}
// x_type <= y_type <= t_type or x_type >= y_type >= t_type // y_type == t_type or x_type <= y_type or x_type >= y_type >= t_type
bool TwoCastEliminater::CheckTypesIsIncrementalOrDecreasing() { bool TwoCastEliminater::CheckTypesIsIncreasingOrDecreasing() {
auto x_type = x_->Type(); auto x_type = x_->Type();
if (x_type->isa<TensorType>()) { if (x_type->isa<TensorType>()) {
x_type = x_type->cast<TensorTypePtr>()->element(); x_type = x_type->cast<TensorTypePtr>()->element();
@ -128,18 +128,21 @@ bool TwoCastEliminater::CheckTypesIsIncrementalOrDecreasing() {
auto x_type_id = x_type->type_id(); auto x_type_id = x_type->type_id();
auto y_type_id = y_type->type_id(); auto y_type_id = y_type->type_id();
auto t_type_id = t_type->type_id(); auto t_type_id = t_type->type_id();
// y_type == t_type
if (y_type_id == t_type_id) { if (y_type_id == t_type_id) {
return true; return true;
} }
// If the precision is incremental or decreasing, the cast can be eliminated. // If the precision is increasing or decreasing, the cast can be eliminated.
// x_type <= y_type // x_type <= y_type
bool incremental = CheckTwoTypes(int_map_, x_type_id, y_type_id) || CheckTwoTypes(uint_map_, x_type_id, y_type_id) || bool increasing = CheckTwoTypes(int_map_, x_type_id, y_type_id) || CheckTwoTypes(uint_map_, x_type_id, y_type_id) ||
CheckTwoTypes(float_map_, x_type_id, y_type_id); CheckTwoTypes(float_map_, x_type_id, y_type_id);
if (increasing) {
return true;
}
// x_type >= y_type >= t_type // x_type >= y_type >= t_type
bool decreasing = CheckThreeTypes(int_map_, t_type_id, y_type_id, x_type_id) || return CheckThreeTypes(int_map_, t_type_id, y_type_id, x_type_id) ||
CheckThreeTypes(uint_map_, t_type_id, y_type_id, x_type_id) || CheckThreeTypes(uint_map_, t_type_id, y_type_id, x_type_id) ||
CheckThreeTypes(float_map_, t_type_id, y_type_id, x_type_id); CheckThreeTypes(float_map_, t_type_id, y_type_id, x_type_id);
return incremental || decreasing;
} }
// {prim::kPrimCast, {prim::kPrimCast, X, Y}, T} // {prim::kPrimCast, {prim::kPrimCast, X, Y}, T}
@ -150,7 +153,7 @@ AnfNodePtr TwoCastEliminater::operator()(const OptimizerPtr &, const AnfNodePtr
if (x_ == nullptr || t_ == nullptr || y_ == nullptr) { if (x_ == nullptr || t_ == nullptr || y_ == nullptr) {
return nullptr; return nullptr;
} }
if (CheckTypesIsIncrementalOrDecreasing()) { if (CheckTypesIsIncreasingOrDecreasing()) {
auto cast_op = python_adapter::GetPyFn("mindspore.ops.operations", "Cast")(); auto cast_op = python_adapter::GetPyFn("mindspore.ops.operations", "Cast")();
ValuePtr cast = parse::data_converter::PyDataToValue(cast_op); ValuePtr cast = parse::data_converter::PyDataToValue(cast_op);
auto cnode = NewCNode({NewValueNode(cast), x_, t_}, node->func_graph()); auto cnode = NewCNode({NewValueNode(cast), x_, t_}, node->func_graph());

6
mindspore/ccsrc/frontend/optimizer/irpass/cast_eliminate.h
View File

@ -51,9 +51,9 @@ class TwoCastEliminater : public AnfVisitor {
} }
private: private:
bool CheckTypesIsIncrementalOrDecreasing(); bool CheckTypesIsIncreasingOrDecreasing();
bool CheckTwoTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2); bool CheckTwoTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2) const;
bool CheckThreeTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2, TypeId type3); bool CheckThreeTypes(const std::map<TypeId, int> &type_map, TypeId type1, TypeId type2, TypeId type3) const;
std::map<TypeId, int> int_map_ = { std::map<TypeId, int> int_map_ = {
{kNumberTypeInt, 0}, {kNumberTypeInt8, 1}, {kNumberTypeInt16, 2}, {kNumberTypeInt32, 3}, {kNumberTypeInt64, 4}}; {kNumberTypeInt, 0}, {kNumberTypeInt8, 1}, {kNumberTypeInt16, 2}, {kNumberTypeInt32, 3}, {kNumberTypeInt64, 4}};
std::map<TypeId, int> uint_map_ = {{kNumberTypeUInt, 0}, std::map<TypeId, int> uint_map_ = {{kNumberTypeUInt, 0},

5
mindspore/ccsrc/plugin/device/cpu/kernel/mkldnn/gru_cpu_kernel.cc
View File

@ -1,5 +1,5 @@
/** /**
* Copyright 2022 Huawei Technologies Co., Ltd * Copyright 2022-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -174,7 +174,8 @@ bool GRUCpuKernelMod::Launch(const std::vector<kernel::AddressPtr> &inputs, cons
reinterpret_cast<float *>(inputs[kInputWeightIndex]->addr) + weight_size_ + weight_h_size_); reinterpret_cast<float *>(inputs[kInputWeightIndex]->addr) + weight_size_ + weight_h_size_);
} else { } else {
auto size = GetSize(bias_desc_); auto size = GetSize(bias_desc_);
if (memset_s(GetDataHandle(bias_memory_), size, 0, size)) { auto ret = memset_s(GetDataHandle(bias_memory_), size, 0, size);
if (ret != EOK) {
MS_LOG(EXCEPTION) << "Bias memset error"; MS_LOG(EXCEPTION) << "Bias memset error";
} }
} }

4
mindspore/ccsrc/plugin/device/cpu/optimizer/print_value_type.h
View File

@ -1,5 +1,5 @@
/** /**
* Copyright 2022 Huawei Technologies Co., Ltd * Copyright 2022-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -23,7 +23,7 @@ namespace mindspore {
namespace opt { namespace opt {
class PrintValueType : public Pass { class PrintValueType : public Pass {
public: public:
explicit PrintValueType(const std::string &name) : Pass("print_value_type") {} explicit PrintValueType(const std::string &) : Pass("print_value_type") {}
~PrintValueType() override = default; ~PrintValueType() override = default;
bool Run(const FuncGraphPtr &graph) override; bool Run(const FuncGraphPtr &graph) override;
}; };

7
mindspore/core/ir/map_tensor.h
View File

@ -1,5 +1,5 @@
/** /**
* Copyright 2022 Huawei Technologies Co., Ltd * Copyright 2022-2023 Huawei Technologies Co., Ltd
* *
* Licensed under the Apache License, Version 2.0 (the "License"); * Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License. * you may not use this file except in compliance with the License.
@ -71,6 +71,7 @@ class MS_CORE_API MapTensor final : public Tensor {
key_shape_ = {abstract::Shape::kShapeDimAny}; key_shape_ = {abstract::Shape::kShapeDimAny};
shape_ = {abstract::Shape::kShapeDimAny}; shape_ = {abstract::Shape::kShapeDimAny};
(void)shape_.insert(shape_.cend(), value_shape.cbegin(), value_shape.cend()); (void)shape_.insert(shape_.cend(), value_shape.cbegin(), value_shape.cend());
size_ = shape_[0];
ShapeVector key_shape = {abstract::Shape::kShapeDimAny}; ShapeVector key_shape = {abstract::Shape::kShapeDimAny};
key_tensor_ = std::make_shared<Tensor>(key_dtype, key_shape); key_tensor_ = std::make_shared<Tensor>(key_dtype, key_shape);
value_tensor_ = std::make_shared<Tensor>(value_dtype, shape_); value_tensor_ = std::make_shared<Tensor>(value_dtype, shape_);
@ -90,13 +91,13 @@ class MS_CORE_API MapTensor final : public Tensor {
MapTensor(const TensorPtr &key_tensor, const TensorPtr &value_tensor, const TensorPtr &status_tensor, MapTensor(const TensorPtr &key_tensor, const TensorPtr &value_tensor, const TensorPtr &status_tensor,
const ValuePtr &default_value, const ValuePtr &permit_filter_value = nullptr, const ValuePtr &default_value, const ValuePtr &permit_filter_value = nullptr,
const ValuePtr &evict_filter_value = nullptr) const ValuePtr &evict_filter_value = nullptr)
: default_value_(default_value) { : key_dtype_(key_tensor->data_type()), default_value_(default_value) {
key_dtype_ = key_tensor->data_type();
data_type_ = value_tensor->data_type(); data_type_ = value_tensor->data_type();
shape_ = value_tensor->shape(); shape_ = value_tensor->shape();
key_shape_ = key_tensor->shape(); key_shape_ = key_tensor->shape();
value_shape_.clear(); value_shape_.clear();
(void)value_shape_.insert(value_shape_.cend(), shape_.cbegin() + 1, shape_.cend()); (void)value_shape_.insert(value_shape_.cend(), shape_.cbegin() + 1, shape_.cend());
size_ = shape_.size() != 0 ? shape_[0] : (abstract::Shape::kShapeDimAny);
key_tensor_ = key_tensor; key_tensor_ = key_tensor;
value_tensor_ = value_tensor; value_tensor_ = value_tensor;
status_tensor_ = status_tensor; status_tensor_ = status_tensor;