[clang-tidy] enable modernize-use-override

.clang-tidy

@@ -185,7 +185,7 @@ Checks: '
 -modernize-use-equals-delete,
 -modernize-use-noexcept,
 -modernize-use-nullptr,
--modernize-use-override,
+modernize-use-override,
 -modernize-use-transparent-functors,
 -modernize-use-uncaught-exceptions,
 -performance-faster-string-find,

paddle/fluid/framework/ir/coalesce_grad_tensor_pass.cc

@@ -69,7 +69,7 @@ double GetFuseParameterMemorySize() { return FLAGS_fuse_parameter_memory_size; }
 
 class CoalesceGradTensorPass : public ir::Pass {
  protected:
-  void ApplyImpl(ir::Graph *graph) const {
+  void ApplyImpl(ir::Graph *graph) const override {
     if (Get<size_t>(details::kNRanks) <= 1) {
       VLOG(6) << "The number of place is" << Get<size_t>(details::kNRanks)
               << ", there doesn't need apply FuseAllReduceOpPass.";

paddle/fluid/framework/ir/fuse_optimizer_ops_pass/fuse_adam_op_pass.cc

@@ -30,9 +30,9 @@ class Node;
 
 class FuseAdamOpPass : public FuseOptimizerOpPass {
  private:
-  const std::string GetOpType() const { return "adam"; }
+  const std::string GetOpType() const override { return "adam"; }
 
-  const std::vector<std::string> GetAuxiliaryVarNames() const {
+  const std::vector<std::string> GetAuxiliaryVarNames() const override {
     return {"Moment1", "Moment2", "Beta1Pow", "Beta2Pow"};
   }
 
@@ -41,7 +41,7 @@ class FuseAdamOpPass : public FuseOptimizerOpPass {
           &aux_var_set,
       const std::unordered_map<std::string, std::string> &fused_vars_name,
       const std::vector<ir::Node *> &adam_ops,
-      ir::Graph *graph) const {
+      ir::Graph *graph) const override {
     auto fused_adam_node =
         FuseAdamOps(aux_var_set, fused_vars_name, adam_ops, graph);
     return fused_adam_node;

paddle/fluid/framework/ir/fuse_optimizer_ops_pass/fuse_momentum_op_pass.cc

@@ -29,18 +29,18 @@ class Node;
 
 class FuseMomentumOpPass : public FuseOptimizerOpPass {
  private:
-  virtual const std::string GetOpType() const { return "momentum"; }
+  const std::string GetOpType() const override { return "momentum"; }
 
-  virtual const std::vector<std::string> GetAuxiliaryVarNames() const {
+  const std::vector<std::string> GetAuxiliaryVarNames() const override {
     return {"Velocity"};
   }
 
   // Fuse Momentum Ops
-  virtual ir::Node *FuseOptimizerOps(
+  ir::Node *FuseOptimizerOps(
       const std::unordered_map<std::string, std::vector<std::string>> &vars_set,
       const std::unordered_map<std::string, std::string> &fused_vars_name,
       const std::vector<ir::Node *> &momentum_ops,
-      ir::Graph *graph) const {
+      ir::Graph *graph) const override {
     PADDLE_ENFORCE_GT(
         momentum_ops.size(),
         static_cast<size_t>(0),

paddle/fluid/framework/ir/fuse_optimizer_ops_pass/fuse_sgd_op_pass.cc

@@ -28,18 +28,18 @@ class Node;
 
 class FuseSgdOpPass : public FuseOptimizerOpPass {
  private:
-  virtual const std::string GetOpType() const { return "sgd"; }
+  const std::string GetOpType() const override { return "sgd"; }
 
-  virtual const std::vector<std::string> GetAuxiliaryVarNames() const {
+  const std::vector<std::string> GetAuxiliaryVarNames() const override {
     return {};
   }
 
   // Fuse Sgd Ops
-  virtual ir::Node *FuseOptimizerOps(
+  ir::Node *FuseOptimizerOps(
       const std::unordered_map<std::string, std::vector<std::string>> &vars_set,
       const std::unordered_map<std::string, std::string> &fused_vars_name,
       const std::vector<ir::Node *> &sgd_ops,
-      ir::Graph *graph) const {
+      ir::Graph *graph) const override {
     PADDLE_ENFORCE_GT(
         sgd_ops.size(),
         static_cast<size_t>(0),

paddle/fluid/framework/new_executor/workqueue/workqueue.cc

@@ -53,7 +53,7 @@ class WorkQueueImpl : public WorkQueue {
                                        options_.always_spinning);
   }
 
-  virtual ~WorkQueueImpl() {
+  ~WorkQueueImpl() override {
     delete queue_;
     if (tracker_ != nullptr) {
       tracker_->~TaskTracker();
@@ -94,7 +94,7 @@ class WorkQueueGroupImpl : public WorkQueueGroup {
   explicit WorkQueueGroupImpl(
       const std::vector<WorkQueueOptions>& queue_options);
 
-  ~WorkQueueGroupImpl();
+  ~WorkQueueGroupImpl() override;
 
   void AddTask(size_t queue_idx, std::function<void()> fn) override;
 

paddle/fluid/framework/phi_utils.cc

@@ -40,7 +40,7 @@ class KernelArgsNameMakerByOpProto : public KernelArgsNameMaker {
         platform::errors::InvalidArgument("Op proto cannot be nullptr."));
   }
 
-  ~KernelArgsNameMakerByOpProto() {}
+  ~KernelArgsNameMakerByOpProto() override {}
 
   const paddle::small_vector<const char*>& GetInputArgsNames() override;
   const paddle::small_vector<const char*>& GetOutputArgsNames() override;

paddle/fluid/memory/allocation/allocator_facade.cc

@@ -127,22 +127,22 @@ class CUDAGraphAllocator
       : underlying_allocator_(allocator) {}
 
  public:
-  ~CUDAGraphAllocator() {}
+  ~CUDAGraphAllocator() override {}
 
   static std::shared_ptr<Allocator> Create(
       const std::shared_ptr<Allocator>& allocator) {
     return std::shared_ptr<Allocator>(new CUDAGraphAllocator(allocator));
   }
 
  protected:
-  phi::Allocation* AllocateImpl(size_t size) {
+  phi::Allocation* AllocateImpl(size_t size) override {
     VLOG(10) << "Allocate " << size << " for CUDA Graph";
     return new PrivateAllocation(this,
                                  static_unique_ptr_cast<Allocation>(
                                      underlying_allocator_->Allocate(size)));
   }
 
-  void FreeImpl(phi::Allocation* allocation) {
+  void FreeImpl(phi::Allocation* allocation) override {
     VLOG(10) << "delete for CUDA Graph";
     delete allocation;
   }

paddle/fluid/operators/collective/allreduce_op.cc

@@ -36,7 +36,7 @@ class AllReduceDelOp : public framework::OperatorWithKernel {
 
 class AllReduceDelOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor), tensor to be allreduced.");
     AddOutput("Out", "(Tensor) the result of allreduced.");
     AddAttr<int>("reduce_type", "(int) determine the reduce type.")

paddle/fluid/operators/collective/alltoall_op.cc

@@ -45,7 +45,7 @@ class AllToAllBaseOp : public framework::OperatorWithKernel {
 
 class AllToAllBaseOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor send.");
     AddOutput("Out", "(Tensor) the result of alltoall.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")

paddle/fluid/operators/collective/barrier_op.cc

@@ -27,7 +27,7 @@ class BarrierOp : public framework::OperatorWithKernel {
 
 class BarrierOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) Input data (only used in CUDAKernel).");
     AddOutput("Out", "(Tensor) Output data (only used in CUDAKernel).");
     AddAttr<int>("ring_id", "(int default 0) communication ring id.")

paddle/fluid/operators/collective/c_allgather_op.cc

@@ -44,7 +44,7 @@ class CAllGatherOp : public framework::OperatorWithKernel {
 
 class CAllGatherOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be allgather");
     AddOutput("Out", "(Tensor) the allgather result");
     AddAttr<int>("ring_id", "(int default 0) communication ring id.")

paddle/fluid/operators/collective/c_broadcast_op.cc

@@ -35,7 +35,7 @@ class CBroadcastOp : public framework::OperatorWithKernel {
 
 class CBroadcastOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be broadcasted.");
     AddOutput("Out", "(Tensor) the result of broadcast.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")

paddle/fluid/operators/collective/c_comm_init_all_op.cc

@@ -38,7 +38,7 @@ namespace operators {
 
 class CCommInitAllInferShape : public framework::InferShapeBase {
  public:
-  ~CCommInitAllInferShape() {}
+  ~CCommInitAllInferShape() override {}
   void operator()(framework::InferShapeContext* ctx) const override{};
 };
 

paddle/fluid/operators/collective/c_comm_init_multitrainer_op.cc

@@ -38,7 +38,7 @@ namespace operators {
 
 class CCommInitMultiTrainerInferShape : public framework::InferShapeBase {
  public:
-  ~CCommInitMultiTrainerInferShape() {}
+  ~CCommInitMultiTrainerInferShape() override {}
   void operator()(framework::InferShapeContext* ctx) const override{};
 };
 

paddle/fluid/operators/collective/c_concat_op.cc

@@ -81,7 +81,7 @@ class CConcatOpGradMaker : public framework::SingleGradOpMaker<T> {
 
 class CConcatOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be concated.");
     AddOutput("Out", "(Tensor) the result of concat.");
     AddAttr<int>("rank", "(int default 0) rank id.").SetDefault(0);

paddle/fluid/operators/collective/c_identity_op.cc

@@ -44,7 +44,7 @@ class CIdentityOp : public framework::OperatorWithKernel {
 
 class CIdentityOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) identity tensor.");
     AddOutput("Out", "(Tensor) identity tensor.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")

paddle/fluid/operators/collective/c_reducescatter_op.cc

@@ -42,7 +42,7 @@ class CReduceScatterOp : public framework::OperatorWithKernel {
 
 class CReduceScatterOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be allgather");
     AddOutput("Out", "(Tensor) the allgather result");
     AddAttr<int>("ring_id", "(int default 0) communication ring id.")

paddle/fluid/operators/collective/c_scatter_op.cc

@@ -61,7 +61,7 @@ class CScatterOp : public framework::OperatorWithKernel {
 
 class CScatterOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be broadcasted.");
     AddOutput("Out", "(Tensor) the result of broadcast.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")

paddle/fluid/operators/collective/c_softmax_with_cross_entropy_op.cc

@@ -83,7 +83,7 @@ class CSoftmaxWithCrossEntropyOp : public framework::OperatorWithKernel {
 class CSoftmaxWithCrossEntropyOpMaker
     : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("Logits",
              "(Tensor, default: Tensor<float>), The input tensor of unscaled "
              "log probabilities, whose dimension :attr:`axis` should be scaled "

paddle/fluid/operators/collective/c_split_op.cc

@@ -89,7 +89,7 @@ class CSplitOpGradMaker : public framework::SingleGradOpMaker<T> {
 
 class CSplitOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be split.");
     AddOutput("Out", "(Tensor) the result of split.");
     AddAttr<int>("rank", "(int default 0) rank id.").SetDefault(0);

paddle/fluid/operators/collective/c_sync_calc_stream_op.cc

@@ -18,7 +18,7 @@ namespace operators {
 
 class CSyncCalcStreamOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) Dependency of the variable need to sync");
     AddOutput("Out", "(Tensor) Dependency of the variable need to sync");
     AddComment(R"DOC(

paddle/fluid/operators/collective/c_sync_comm_stream_op.cc

@@ -31,7 +31,7 @@ class CSyncCommStreamOp : public framework::OperatorWithKernel {
 
 class CSyncCommStreamOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) Dependency of the variable need to sync")
         .AsDuplicable();
     AddOutput("Out", "(Tensor) Dependency of the variable need to sync")

paddle/fluid/operators/collective/c_wait_comm_op.cc

@@ -73,7 +73,7 @@ class CWaitCommOp : public framework::OperatorBase {
 
 class CWaitCommOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) Dependency of the variable need to sync")
         .AsDuplicable();
     AddOutput("Out", "(Tensor) Dependency of the variable need to sync")

paddle/fluid/operators/collective/c_wait_compute_op.cc

@@ -74,7 +74,7 @@ class CWaitComputeOp : public framework::OperatorBase {
 
 class CWaitComputeOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) Dependency of the variable need to sync")
         .AsDuplicable();
     AddOutput("Out", "(Tensor) Dependency of the variable need to sync")

paddle/fluid/operators/collective/global_gather_op.cc

@@ -58,7 +58,7 @@ class GlobalGatherOp : public framework::OperatorWithKernel {
 
 class GlobalGatherOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor send.");
     AddInput("local_count",
              "(Tensor) Tensor which has n_expert * world_size elements that "

paddle/fluid/operators/collective/global_scatter_op.cc

@@ -61,7 +61,7 @@ class GlobalScatterOp : public framework::OperatorWithKernel {
 
 class GlobalScatterOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor send.");
     AddInput("local_count",
              "(Tensor) Tensor which has n_expert * world_size elements that "

paddle/fluid/operators/collective/mp_allreduce_sum_op.cc

@@ -37,7 +37,7 @@ class MpAllReduceSumOp : public framework::OperatorWithKernel {
 
 class MpAllReduceSumOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor), tensor to be allreduced in model parallel.");
     AddOutput("Out", "(Tensor) the allreduced result in model parallel.");
     AddAttr<int>("ring_id", "(int default 0) communication ring id.")

paddle/fluid/operators/collective/partial_allgather_op.cc

@@ -45,7 +45,7 @@ class PartialAllGatherOp : public framework::OperatorWithKernel {
 
 class PartialAllGatherOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be partial allgather");
     AddOutput("Out", "(Tensor) the allgather result");
     AddAttr<int>("ring_id", "(int default 0) communication ring id.")

paddle/fluid/operators/collective/partial_recv_op.cc

@@ -91,7 +91,7 @@ class PartialRecvOp : public framework::OperatorWithKernel {
 
 class PartialRecvOpMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddOutput("Out", "(Tensor) tensor to receive.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")
         .SetDefault(0);

paddle/fluid/operators/collective/partial_send_op.cc

@@ -60,7 +60,7 @@ class PartialSendOp : public framework::OperatorWithKernel {
 
 class PartialSendMaker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be sent.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")
         .SetDefault(0);

paddle/fluid/operators/collective/recv_v2_op.cc

@@ -80,7 +80,7 @@ class RecvOpV2 : public framework::OperatorWithKernel {
 
 class RecvOpV2Maker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddOutput("Out", "(Tensor) tensor to receive.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")
         .SetDefault(0);

paddle/fluid/operators/collective/send_v2_op.cc

@@ -56,7 +56,7 @@ class SendOpV2 : public framework::OperatorWithKernel {
 
 class SendOpV2Maker : public framework::OpProtoAndCheckerMaker {
  public:
-  void Make() {
+  void Make() override {
     AddInput("X", "(Tensor) tensor to be sent.");
     AddAttr<int>("ring_id", "(int default 0) nccl communication ring id.")
         .SetDefault(0);

paddle/fluid/operators/fused/fused_gemm_epilogue_op.cc

@@ -127,7 +127,7 @@ class FusedGemmEpilogueOp : public framework::OperatorWithKernel {
   }
 
   phi::KernelKey GetExpectedKernelType(
-      const framework::ExecutionContext& ctx) const {
+      const framework::ExecutionContext& ctx) const override {
     auto data_type = OperatorWithKernel::IndicateVarDataType(ctx, "X");
     return phi::KernelKey(data_type, ctx.GetPlace());
   }
@@ -277,7 +277,7 @@ class FusedGemmEpilogueGradOp : public framework::OperatorWithKernel {
   }
 
   phi::KernelKey GetExpectedKernelType(
-      const framework::ExecutionContext& ctx) const {
+      const framework::ExecutionContext& ctx) const override {
     auto data_type = OperatorWithKernel::IndicateVarDataType(ctx, "DOut");
     return phi::KernelKey(data_type, ctx.GetPlace());
   }