diff --git a/paddle/fluid/distributed/ps/service/brpc_utils.cc b/paddle/fluid/distributed/ps/service/brpc_utils.cc
index 53edb62ff83fc..f07b66e1e90b9 100644
--- a/paddle/fluid/distributed/ps/service/brpc_utils.cc
+++ b/paddle/fluid/distributed/ps/service/brpc_utils.cc
@@ -111,14 +111,13 @@ void SerializeLodTensor(framework::Variable* var,
   }
   // IO Buffer
   if (platform::is_cpu_place(tensor->place())) {
-    auto data_len = tensor->numel() * framework::DataTypeSize(tensor->dtype());
+    auto data_len = tensor->numel() * phi::SizeOf(tensor->dtype());
     iobuf->append(reinterpret_cast<const char*>(&data_len), 8);
     iobuf->append(reinterpret_cast<const char*>(tensor->data()), data_len);
   } else {
 #ifdef PADDLE_WITH_CUDA
     char* temp_ptr =
-        new char[tensor->numel() *
-                 framework::DataTypeSize(tensor->dtype())];  // NOLINT
+        new char[tensor->numel() * phi::SizeOf(tensor->dtype())];  // NOLINT
     auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
     memory::Copy(
         platform::CPUPlace(),
@@ -128,7 +127,7 @@ void SerializeLodTensor(framework::Variable* var,
         tensor->numel() * framework::SizeOfType(
                               framework::TransToProtoVarType(tensor->dtype())),
         stream);
-    auto data_len = tensor->numel() * framework::DataTypeSize(tensor->dtype());
+    auto data_len = tensor->numel() * phi::SizeOf(tensor->dtype());
     iobuf->append(reinterpret_cast<const char*>(&data_len), 8);
     iobuf->append(reinterpret_cast<const char*>(temp_ptr), data_len);
     delete[] temp_ptr;
@@ -159,14 +158,13 @@ void SerializeSelectedRows(framework::Variable* var,
   }
   // IO Buffer
   if (platform::is_cpu_place(tensor->place())) {
-    auto data_len = tensor->numel() * framework::DataTypeSize(tensor->dtype());
+    auto data_len = tensor->numel() * phi::SizeOf(tensor->dtype());
     iobuf->append(reinterpret_cast<const char*>(&data_len), 8);
     iobuf->append(reinterpret_cast<const char*>(tensor->data()), data_len);
   } else {
 #ifdef PADDLE_WITH_CUDA
     char* temp_ptr =
-        new char[tensor->numel() *
-                 framework::DataTypeSize(tensor->dtype())];  // NOLINT
+        new char[tensor->numel() * phi::SizeOf(tensor->dtype())];  // NOLINT
     auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
     memory::Copy(
         platform::CPUPlace(),
@@ -176,7 +174,7 @@ void SerializeSelectedRows(framework::Variable* var,
         tensor->numel() * framework::SizeOfType(
                               framework::TransToProtoVarType(tensor->dtype())),
         stream);
-    auto data_len = tensor->numel() * framework::DataTypeSize(tensor->dtype());
+    auto data_len = tensor->numel() * phi::SizeOf(tensor->dtype());
     iobuf->append(reinterpret_cast<const char*>(&data_len), 8);
     iobuf->append(reinterpret_cast<const char*>(temp_ptr), data_len);
     delete[] temp_ptr;
@@ -259,16 +257,15 @@ void DeserializeLodTensor(framework::Variable* var,
 #ifdef PADDLE_WITH_CUDA
     unsigned long data_len;  // NOLINT
     char* temp_ptr =
-        new char[tensor->numel() *
-                 framework::DataTypeSize(tensor->dtype())];     // NOLINT
-    io_buffer_itr.copy_and_forward((void*)(&data_len), 8);      // NOLINT
-    io_buffer_itr.copy_and_forward((void*)temp_ptr, data_len);  // NOLINT
+        new char[tensor->numel() * phi::SizeOf(tensor->dtype())];  // NOLINT
+    io_buffer_itr.copy_and_forward((void*)(&data_len), 8);         // NOLINT
+    io_buffer_itr.copy_and_forward((void*)temp_ptr, data_len);     // NOLINT
     auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
     memory::Copy(place,
                  tensor_data,
                  platform::CPUPlace(),
                  (void*)temp_ptr,  // NOLINT
-                 tensor->numel() * framework::DataTypeSize(tensor->dtype()),
+                 tensor->numel() * phi::SizeOf(tensor->dtype()),
                  stream);
     delete[] temp_ptr;
 #endif
@@ -303,17 +300,16 @@ void DeserializeSelectedRows(
   } else if (platform::is_gpu_place(place)) {
 #ifdef PADDLE_WITH_CUDA
     char* temp_ptr =
-        new char[tensor->numel() *
-                 framework::DataTypeSize(tensor->dtype())];  // NOLINT
-    unsigned long data_len;                                  // NOLINT
-    io_buffer_itr.copy_and_forward((void*)(&data_len), 8);   // NOLINT
+        new char[tensor->numel() * phi::SizeOf(tensor->dtype())];  // NOLINT
+    unsigned long data_len;                                        // NOLINT
+    io_buffer_itr.copy_and_forward((void*)(&data_len), 8);         // NOLINT
     io_buffer_itr.copy_and_forward(temp_ptr, data_len);
     auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
     memory::Copy(place,
                  tensor_data,
                  platform::CPUPlace(),
                  temp_ptr,
-                 tensor->numel() * framework::DataTypeSize(tensor->dtype()),
+                 tensor->numel() * phi::SizeOf(tensor->dtype()),
                  stream);
     delete[] temp_ptr;
 #endif
diff --git a/paddle/fluid/distributed/ps/service/heter_client.cc b/paddle/fluid/distributed/ps/service/heter_client.cc
index f90893bc8dc2d..2ca9fef5c0876 100644
--- a/paddle/fluid/distributed/ps/service/heter_client.cc
+++ b/paddle/fluid/distributed/ps/service/heter_client.cc
@@ -41,14 +41,14 @@ int GetMicroId(const platform::DeviceContext& ctx,
   } else {
 #ifdef PADDLE_WITH_CUDA
     std::vector<char> temp;
-    temp.resize(tensor->numel() * framework::DataTypeSize(tensor->dtype()));
+    temp.resize(tensor->numel() * phi::SizeOf(tensor->dtype()));
     char* temp_ptr = temp.data();
     auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
     memory::Copy(platform::CPUPlace(),
                  temp_ptr,
                  tensor->place(),
                  tensor->data(),
-                 tensor->numel() * framework::DataTypeSize(tensor->dtype()),
+                 tensor->numel() * phi::SizeOf(tensor->dtype()),
                  stream);
     float* temp_ptr_float = reinterpret_cast<float*>(temp_ptr);
     micro_id = static_cast<int>(temp_ptr_float[0]);
diff --git a/paddle/fluid/eager/amp_auto_cast.h b/paddle/fluid/eager/amp_auto_cast.h
index a16dd95396427..0bf5875823bd9 100644
--- a/paddle/fluid/eager/amp_auto_cast.h
+++ b/paddle/fluid/eager/amp_auto_cast.h
@@ -49,7 +49,7 @@ inline std::vector<paddle::experimental::Tensor> AmpAutoCasts(
     std::string op_name) {
   VLOG(6) << "AMP AmpAutoCasts:"
           << " inputs(" << inputs_name << ") dst_dtype("
-          << paddle::framework::DataType2String(dst_dtype) << ").";
+          << phi::DataTypeToString(dst_dtype) << ").";
   std::vector<paddle::experimental::Tensor> inputs_casted;
   for (auto& input : inputs) {
     if (NeedCast(input, dst_dtype)) {
@@ -72,7 +72,7 @@ inline paddle::experimental::Tensor AmpAutoCast(
     std::string op_name) {
   VLOG(6) << "AMP AmpAutoCasts:"
           << " input(" << input_name << ") dst_dtype("
-          << paddle::framework::DataType2String(dst_dtype) << ").";
+          << phi::DataTypeToString(dst_dtype) << ").";
   if (dst_dtype == paddle::experimental::DataType::FLOAT16) {
     if (op_name == "run_program") {
       return input;
diff --git a/paddle/fluid/eager/eager_amp_auto_cast.h b/paddle/fluid/eager/eager_amp_auto_cast.h
index 22748e31cfd7a..ea3e53b972d99 100644
--- a/paddle/fluid/eager/eager_amp_auto_cast.h
+++ b/paddle/fluid/eager/eager_amp_auto_cast.h
@@ -69,7 +69,7 @@ inline std::vector<paddle::experimental::Tensor> EagerAmpAutoCasts(
     bool trace_backward = true) {
   VLOG(6) << "AMP AmpAutoCasts:"
           << " inputs(" << inputs_name << ") dst_dtype("
-          << paddle::framework::DataType2String(dst_dtype) << ").";
+          << phi::DataTypeToString(dst_dtype) << ").";
   std::vector<paddle::experimental::Tensor> inputs_casted;
   for (auto& input : inputs) {
     if (NeedCast(input, dst_dtype)) {
@@ -89,7 +89,7 @@ inline paddle::experimental::Tensor EagerAmpAutoCast(
     bool trace_backward = true) {
   VLOG(6) << "AMP AmpAutoCasts:"
           << " input(" << egr::EagerUtils::TensorStr(input) << " to dst_dtype("
-          << paddle::framework::DataType2String(dst_dtype) << ").";
+          << phi::DataTypeToString(dst_dtype) << ").";
   if (dst_dtype == paddle::experimental::DataType::FLOAT16) {
     if (op_name == "run_program") {
       return input;
diff --git a/paddle/fluid/eager/grad_node_info.cc b/paddle/fluid/eager/grad_node_info.cc
index afa8a6f205259..43da47436e2ac 100644
--- a/paddle/fluid/eager/grad_node_info.cc
+++ b/paddle/fluid/eager/grad_node_info.cc
@@ -41,16 +41,16 @@ static void CheckTensor(const paddle::experimental::Tensor& pre,
         "The tensor in before and after hook are not consistent"));
   }
   if (pre.initialized() && post.initialized()) {
-    VLOG(7) << paddle::framework::DataType2String(pre.dtype()) << " "
-            << paddle::framework::DataType2String(post.dtype());
+    VLOG(7) << phi::DataTypeToString(pre.dtype()) << " "
+            << phi::DataTypeToString(post.dtype());
     PADDLE_ENFORCE_EQ(
         pre.dtype(),
         post.dtype(),
         paddle::platform::errors::PermissionDenied(
             "The dtype of tensor before(%s) and after(%s) hook are not "
             "consistent",
-            paddle::framework::DataType2String(pre.dtype()),
-            paddle::framework::DataType2String(post.dtype())));
+            phi::DataTypeToString(pre.dtype()),
+            phi::DataTypeToString(post.dtype())));
     PADDLE_ENFORCE_EQ(pre.place(),
                       post.place(),
                       paddle::platform::errors::PermissionDenied(
diff --git a/paddle/fluid/framework/CMakeLists.txt b/paddle/fluid/framework/CMakeLists.txt
index 6fd4095d0d28f..05c2a4ccfb2c4 100755
--- a/paddle/fluid/framework/CMakeLists.txt
+++ b/paddle/fluid/framework/CMakeLists.txt
@@ -1242,23 +1242,7 @@ cc_test(
   SRCS phi_utils_test.cc
   DEPS phi_utils)
 
-if(WITH_GPU OR WITH_ROCM)
-  cc_library(
-    fluid_convert_utils
-    SRCS convert_utils.cc
-    DEPS data_type place gpu_info)
-else()
-  cc_library(
-    fluid_convert_utils
-    SRCS convert_utils.cc
-    DEPS data_type place)
-endif()
-
-# every source file that includes "dnnl.h" must depends on mkldnn
-# or, the first one should depends on mkldnn
-if(WITH_MKLDNN)
-  add_dependencies(fluid_convert_utils mkldnn)
-endif()
+cc_library(fluid_convert_utils DEPS data_type)
 
 cc_test(
   convert_utils_test
diff --git a/paddle/fluid/framework/convert_utils.cc b/paddle/fluid/framework/convert_utils.cc
deleted file mode 100644
index 49efde53f4a8e..0000000000000
--- a/paddle/fluid/framework/convert_utils.cc
+++ /dev/null
@@ -1,166 +0,0 @@
-/* Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License. */
-
-#include "paddle/fluid/framework/convert_utils.h"
-// See Note [ Why still include the fluid headers? ]
-#include "paddle/fluid/platform/device/gpu/gpu_info.h"
-
-#include "paddle/phi/common/pstring.h"
-namespace paddle {
-namespace framework {
-
-paddle::experimental::DataType TransToPhiDataType(
-    const paddle::framework::proto::VarType::Type& dtype) {
-  // Set the order of case branches according to the frequency with
-  // the data type is used
-  switch (dtype) {
-    case paddle::framework::proto::VarType::FP32:
-      return DataType::FLOAT32;
-    case paddle::framework::proto::VarType::FP64:
-      return DataType::FLOAT64;
-    case paddle::framework::proto::VarType::INT64:
-      return DataType::INT64;
-    case paddle::framework::proto::VarType::INT32:
-      return DataType::INT32;
-    case paddle::framework::proto::VarType::INT8:
-      return DataType::INT8;
-    case paddle::framework::proto::VarType::UINT8:
-      return DataType::UINT8;
-    case paddle::framework::proto::VarType::INT16:
-      return DataType::INT16;
-    case paddle::framework::proto::VarType::COMPLEX64:
-      return DataType::COMPLEX64;
-    case paddle::framework::proto::VarType::COMPLEX128:
-      return DataType::COMPLEX128;
-    case paddle::framework::proto::VarType::FP16:
-      return DataType::FLOAT16;
-    case paddle::framework::proto::VarType::BF16:
-      return DataType::BFLOAT16;
-    case paddle::framework::proto::VarType::BOOL:
-      return DataType::BOOL;
-    case paddle::framework::proto::VarType::PSTRING:
-      return DataType::PSTRING;
-    default:
-      return DataType::UNDEFINED;
-  }
-}
-
-paddle::framework::proto::VarType::Type TransToProtoVarType(
-    const paddle::experimental::DataType& dtype) {
-  // Set the order of case branches according to the frequency with
-  // the data type is used
-  switch (dtype) {
-    case DataType::FLOAT32:
-      return paddle::framework::proto::VarType::FP32;
-    case DataType::FLOAT64:
-      return paddle::framework::proto::VarType::FP64;
-    case DataType::INT64:
-      return paddle::framework::proto::VarType::INT64;
-    case DataType::INT32:
-      return paddle::framework::proto::VarType::INT32;
-    case DataType::INT8:
-      return paddle::framework::proto::VarType::INT8;
-    case DataType::UINT8:
-      return paddle::framework::proto::VarType::UINT8;
-    case DataType::INT16:
-      return paddle::framework::proto::VarType::INT16;
-    case DataType::COMPLEX64:
-      return paddle::framework::proto::VarType::COMPLEX64;
-    case DataType::COMPLEX128:
-      return paddle::framework::proto::VarType::COMPLEX128;
-    case DataType::FLOAT16:
-      return paddle::framework::proto::VarType::FP16;
-    case DataType::BFLOAT16:
-      return paddle::framework::proto::VarType::BF16;
-    case DataType::BOOL:
-      return paddle::framework::proto::VarType::BOOL;
-    case DataType::PSTRING:
-      return paddle::framework::proto::VarType::PSTRING;
-    default:
-      PADDLE_THROW(paddle::platform::errors::Unimplemented(
-          "Unsupported data type `%s` when casting it into "
-          "paddle data type.",
-          dtype));
-  }
-}
-
-size_t DataTypeSize(DataType dtype) {
-  switch (dtype) {
-    case DataType::UNDEFINED:
-      return 0;
-    case DataType::BOOL:
-      return sizeof(bool);
-    case DataType::INT8:
-      return sizeof(int8_t);
-    case DataType::UINT8:
-      return sizeof(uint8_t);
-    case DataType::INT16:
-      return sizeof(int16_t);
-    case DataType::INT32:
-      return sizeof(int);
-    case DataType::INT64:
-      return sizeof(int64_t);
-    case DataType::BFLOAT16:
-      return sizeof(paddle::platform::bfloat16);
-    case DataType::FLOAT16:
-      return sizeof(paddle::platform::float16);
-    case DataType::FLOAT32:
-      return sizeof(float);
-    case DataType::FLOAT64:
-      return sizeof(double);
-    case DataType::COMPLEX64:
-      return sizeof(paddle::platform::complex<float>);
-    case DataType::COMPLEX128:
-      return sizeof(paddle::platform::complex<double>);
-    case DataType::PSTRING:
-      return sizeof(paddle::platform::pstring);
-    default:
-      return 0;
-  }
-}
-
-DataType String2DataType(const std::string& str) {
-  if (str == "bool") {
-    return DataType::BOOL;
-  } else if (str == "float16") {
-    return DataType::FLOAT16;
-  } else if (str == "float32") {
-    return DataType::FLOAT32;
-  } else if (str == "float64") {
-    return DataType::FLOAT64;
-  } else if (str == "int8") {
-    return DataType::INT8;
-  } else if (str == "int16") {
-    return DataType::INT16;
-  } else if (str == "int32") {
-    return DataType::INT32;
-  } else if (str == "int64") {
-    return DataType::INT64;
-  } else if (str == "uint8") {
-    return DataType::UINT8;
-  } else if (str == "complex64") {
-    return DataType::COMPLEX64;
-  } else if (str == "complex128") {
-    return DataType::COMPLEX128;
-  } else if (str == "pstring") {
-    return DataType::PSTRING;
-  } else if (str == "bfloat16") {
-    return DataType::BFLOAT16;
-  } else {
-    return DataType::UNDEFINED;
-  }
-}
-
-}  // namespace framework
-}  // namespace paddle
diff --git a/paddle/fluid/framework/convert_utils.h b/paddle/fluid/framework/convert_utils.h
index d3aca94003045..6995e21da8910 100644
--- a/paddle/fluid/framework/convert_utils.h
+++ b/paddle/fluid/framework/convert_utils.h
@@ -14,12 +14,8 @@ limitations under the License. */
 
 #pragma once
 
-#include "paddle/phi/common/backend.h"
-#include "paddle/phi/common/data_type.h"
-#include "paddle/phi/common/layout.h"
-#include "paddle/phi/core/tensor_meta.h"
-
 #include "paddle/fluid/framework/data_type.h"
+#include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/utils/data_type.h"
 
 // TODO(chenweihang): this file may need to be removed
@@ -27,19 +23,16 @@ limitations under the License. */
 namespace paddle {
 namespace framework {
 
-using DataType = paddle::experimental::DataType;
+using DataType = phi::DataType;
 using DataLayout = phi::DataLayout;
 
-DataType TransToPhiDataType(
-    const paddle::framework::proto::VarType::Type& dtype);
-
-paddle::framework::proto::VarType::Type TransToProtoVarType(
-    const DataType& dtype);
-
-size_t DataTypeSize(DataType dtype);
-DataType String2DataType(const std::string& str);
+using phi::DataTypeToString;
+using phi::SizeOf;
+using phi::TransToPhiDataType;
 
-using phi::DataType2String;
+inline proto::VarType::Type TransToProtoVarType(const DataType& dtype) {
+  return static_cast<proto::VarType::Type>(phi::TransToProtoVarType(dtype));
+}
 
 }  // namespace framework
 }  // namespace paddle
diff --git a/paddle/fluid/framework/heter_section_worker.cc b/paddle/fluid/framework/heter_section_worker.cc
index 914a172039ec2..6321fe9c5388f 100644
--- a/paddle/fluid/framework/heter_section_worker.cc
+++ b/paddle/fluid/framework/heter_section_worker.cc
@@ -45,7 +45,7 @@ void SetMicroId(paddle::framework::Scope* scope,
   if (platform::is_gpu_place(place)) {
 #ifdef PADDLE_WITH_CUDA
     std::vector<char> temp;
-    temp.resize(tensor->numel() * framework::DataTypeSize(tensor->dtype()));
+    temp.resize(tensor->numel() * phi::SizeOf(tensor->dtype()));
     char* temp_ptr = temp.data();
     float* temp_ptr_float = reinterpret_cast<float*>(temp_ptr);
     temp_ptr_float[0] = micro_id;
diff --git a/paddle/fluid/framework/tensor_util.cc b/paddle/fluid/framework/tensor_util.cc
index 867f15a3e09bd..91b87a98447ce 100644
--- a/paddle/fluid/framework/tensor_util.cc
+++ b/paddle/fluid/framework/tensor_util.cc
@@ -74,9 +74,9 @@ void TensorCopyImpl(const TENSOR& src,
 #ifdef PADDLE_WITH_MKLDNN
   auto size = src.layout() == DataLayout::ONEDNN
                   ? src.memory_size()
-                  : src.numel() * framework::DataTypeSize(src.dtype());
+                  : src.numel() * phi::SizeOf(src.dtype());
 #else
-  auto size = src.numel() * framework::DataTypeSize(src.dtype());
+  auto size = src.numel() * phi::SizeOf(src.dtype());
 #endif
 
   if (platform::is_cpu_place(src_place) && platform::is_cpu_place(dst_place)) {
@@ -486,7 +486,7 @@ void TensorCopySync(const phi::DenseTensor& src,
     return;
   }
 
-  auto size = src.numel() * framework::DataTypeSize(src.dtype());
+  auto size = src.numel() * phi::SizeOf(src.dtype());
   if (platform::is_cpu_place(src_place) && platform::is_cpu_place(dst_place)) {
     memory::Copy(dst_place, dst_ptr, src_place, src_ptr, size);
   }
@@ -679,7 +679,7 @@ void TensorToStream(std::ostream& os,
     os.write(out.data(), size);
   }
   {  // the 3rd field, tensor data
-    uint64_t size = tensor.numel() * framework::DataTypeSize(tensor.dtype());
+    uint64_t size = tensor.numel() * phi::SizeOf(tensor.dtype());
 
     auto* data_ptr = tensor.data();
     PADDLE_ENFORCE_LT(size,
diff --git a/paddle/fluid/inference/lite/tensor_utils.cc b/paddle/fluid/inference/lite/tensor_utils.cc
index 3208ef961dcf0..67e81d676a179 100644
--- a/paddle/fluid/inference/lite/tensor_utils.cc
+++ b/paddle/fluid/inference/lite/tensor_utils.cc
@@ -217,7 +217,7 @@ void TensorCopyAsync(paddle::lite_api::Tensor* dst,
   const platform::Place& src_place = src.place();
   const platform::Place& dst_place = GetNativePlace(dst->target());
   const size_t bytes =
-      static_cast<size_t>(src.numel()) * framework::DataTypeSize(src.dtype());
+      static_cast<size_t>(src.numel()) * phi::SizeOf(src.dtype());
   dst->Resize(phi::vectorize(src.dims()));
   const void* src_data = src.data();
   void* dst_data{nullptr};
@@ -241,7 +241,7 @@ void TensorCopyAsync(phi::DenseTensor* dst,
   const platform::Place& src_place = GetNativePlace(src.target());
   const platform::Place& dst_place = dst->place();
   int64_t src_numel = GetLiteTensorNumel(src);
-  const size_t bytes = src_numel * framework::DataTypeSize(dst->dtype());
+  const size_t bytes = src_numel * phi::SizeOf(dst->dtype());
   const void* src_data = src.data<void>();
   // When Lite is ready, the source type needs to be modified here.
   void* dst_data = dst->mutable_data(dst_place, dst->dtype());
diff --git a/paddle/fluid/operators/amp/update_loss_scaling_op_npu.cc b/paddle/fluid/operators/amp/update_loss_scaling_op_npu.cc
index dc1cd958f458c..b1bfcf8edd672 100644
--- a/paddle/fluid/operators/amp/update_loss_scaling_op_npu.cc
+++ b/paddle/fluid/operators/amp/update_loss_scaling_op_npu.cc
@@ -219,7 +219,7 @@ class LazyZerosNPU {
       if (!found_inf_vec[0]) {
         framework::TensorCopy(*x, place, dev_ctx, out);
       } else if (zero_ptr != dst_ptr) {
-        auto size = out->numel() * framework::DataTypeSize(out->dtype());
+        auto size = out->numel() * phi::SizeOf(out->dtype());
         memory::Copy(place, dst_ptr, place, zero_ptr, size, stream);
       }
     }
diff --git a/paddle/fluid/operators/collective/c_embedding_op.h b/paddle/fluid/operators/collective/c_embedding_op.h
index 4dbc6f963194f..55fd021a7cde3 100644
--- a/paddle/fluid/operators/collective/c_embedding_op.h
+++ b/paddle/fluid/operators/collective/c_embedding_op.h
@@ -128,7 +128,7 @@ class CEmbeddingGradOpCPUKernel : public framework::OpKernel<T> {
         table_grad_t->mutable_data<T>(table_t->dims(), context.GetPlace());
 
     size_t table_t_mem_size =
-        table_t->numel() * framework::DataTypeSize(table_grad_t->dtype());
+        table_t->numel() * phi::SizeOf(table_grad_t->dtype());
     size_t table_grad_t_mem_size =
         table_grad_t->numel() *
         framework::SizeOfType(
diff --git a/paddle/fluid/operators/collective/c_embedding_op_npu.cc b/paddle/fluid/operators/collective/c_embedding_op_npu.cc
index 263fb264dc10d..f7fd7fbd83eab 100644
--- a/paddle/fluid/operators/collective/c_embedding_op_npu.cc
+++ b/paddle/fluid/operators/collective/c_embedding_op_npu.cc
@@ -127,10 +127,8 @@ void NPUGetIdsEmbedding(const framework::ExecutionContext &context) {
   auto pad_shape = phi::make_ddim({table_t->dims()[0] + 1, table_t->dims()[1]});
   phi::DenseTensor table_t_pad;
 
-  size_t mem_size =
-      table_t->numel() * framework::DataTypeSize(table_t->dtype());
-  size_t line_mem_size =
-      table_t->dims()[1] * framework::DataTypeSize(table_t->dtype());
+  size_t mem_size = table_t->numel() * phi::SizeOf(table_t->dtype());
+  size_t line_mem_size = table_t->dims()[1] * phi::SizeOf(table_t->dtype());
   PADDLE_ENFORCE_EQ(line_mem_size % 64,
                     0,
                     platform::errors::InvalidArgument(
@@ -227,11 +225,11 @@ void NPUUpdateEmbedding(const framework::ExecutionContext &context) {
   // copy table_t_pad to table_t
   T *dst = table_grad_t->mutable_data<T>(table_t->dims(), context.GetPlace());
   const size_t mem_size =
-      table_grad_t->numel() * framework::DataTypeSize(table_grad_t->dtype());
+      table_grad_t->numel() * phi::SizeOf(table_grad_t->dtype());
 
   // check align
   size_t line_mem_size =
-      table_grad_t->dims()[1] * framework::DataTypeSize(table_grad_t->dtype());
+      table_grad_t->dims()[1] * phi::SizeOf(table_grad_t->dtype());
   PADDLE_ENFORCE_EQ(line_mem_size % 64,
                     0,
                     platform::errors::InvalidArgument(
diff --git a/paddle/fluid/operators/detection/bbox_util.h b/paddle/fluid/operators/detection/bbox_util.h
index 4046f6b2830d8..aba166355ed8e 100644
--- a/paddle/fluid/operators/detection/bbox_util.h
+++ b/paddle/fluid/operators/detection/bbox_util.h
@@ -153,7 +153,7 @@ static void AppendProposals(phi::DenseTensor* dst,
                             const phi::DenseTensor& src) {
   auto* out_data = dst->data();
   auto* to_add_data = src.data();
-  size_t size_of_t = framework::DataTypeSize(src.dtype());
+  size_t size_of_t = phi::SizeOf(src.dtype());
   offset *= size_of_t;
   std::memcpy(
       reinterpret_cast<void*>(reinterpret_cast<uintptr_t>(out_data) + offset),
diff --git a/paddle/fluid/operators/ipu/ipu_runtime_op.cc b/paddle/fluid/operators/ipu/ipu_runtime_op.cc
index e243c8f7d9e36..2c7bde2871a25 100644
--- a/paddle/fluid/operators/ipu/ipu_runtime_op.cc
+++ b/paddle/fluid/operators/ipu/ipu_runtime_op.cc
@@ -46,7 +46,7 @@ class IpuRuntimeOp : public framework::OperatorBase {
     for (size_t i = 0; i < outputs.size(); ++i) {
       auto* out = outputs[i];
       if (out->dims().size() == 0) {
-        auto sizeof_dtype = framework::DataTypeSize(out->dtype());
+        auto sizeof_dtype = phi::SizeOf(out->dtype());
         int64_t dim = out->memory_size() / sizeof_dtype;
         out->Resize({dim});
         VLOG(10) << "set ipu_runtime_op output: " << output_names[i]
diff --git a/paddle/fluid/operators/pscore/send_and_recv_op_gpu_test.cc b/paddle/fluid/operators/pscore/send_and_recv_op_gpu_test.cc
index 0491acc059919..faac4865975e5 100644
--- a/paddle/fluid/operators/pscore/send_and_recv_op_gpu_test.cc
+++ b/paddle/fluid/operators/pscore/send_and_recv_op_gpu_test.cc
@@ -128,7 +128,7 @@ void InitTensorsOnClient(framework::Scope* scope,
                reinterpret_cast<void*>(x_ptr),
                platform::CPUPlace(),
                reinterpret_cast<void*>(x_vec_ptr),
-               x_var->numel() * framework::DataTypeSize(x_var->dtype()),
+               x_var->numel() * phi::SizeOf(x_var->dtype()),
                stream);
 
   // auto res_var = scope->Var("res")->GetMutable<phi::DenseTensor>();
diff --git a/paddle/fluid/operators/reader/buffered_reader.cc b/paddle/fluid/operators/reader/buffered_reader.cc
index 8ba7b84fe4bf9..5bb8a29ce356e 100644
--- a/paddle/fluid/operators/reader/buffered_reader.cc
+++ b/paddle/fluid/operators/reader/buffered_reader.cc
@@ -191,8 +191,7 @@ void BufferedReader::ReadAsync(size_t i) {
             cuda[i].set_layout(cpu[i].layout());
             cuda_pinned_ptrs[i] =
                 cuda[i].mutable_data(cuda_pinned_place, cpu[i].type());
-            auto size = cpu[i].numel() *
-                        paddle::framework::DataTypeSize(cpu[i].dtype());
+            auto size = cpu[i].numel() * phi::SizeOf(cpu[i].dtype());
 
             memory::Copy(cuda_pinned_place,
                          cuda_pinned_ptrs[i],
@@ -245,8 +244,7 @@ void BufferedReader::ReadAsync(size_t i) {
           auto cpu_place = cpu[i].place();
           auto cpu_ptr = cpu[i].data();
           auto gpu_ptr = gpu_ptrs[i];
-          auto size =
-              cpu[i].numel() * paddle::framework::DataTypeSize(cpu[i].dtype());
+          auto size = cpu[i].numel() * phi::SizeOf(cpu[i].dtype());
           if (platform::is_cuda_pinned_place(cpu_place)) {
             memory::Copy(
                 place_, gpu_ptr, cpu_place, cpu_ptr, size, stream_.get());
@@ -312,8 +310,7 @@ void BufferedReader::ReadAsync(size_t i) {
         auto cpu_place = cpu[i].place();
         auto cpu_ptr = cpu[i].data();
         auto npu_ptr = npu_ptrs[i];
-        auto size =
-            cpu[i].numel() * paddle::framework::DataTypeSize(cpu[i].dtype());
+        auto size = cpu[i].numel() * phi::SizeOf(cpu[i].dtype());
         if ((platform::is_npu_place(cpu_place))) {
           memory::Copy(
               place_, npu_ptr, cpu_place, cpu_ptr, size, stream_.get());
@@ -364,8 +361,7 @@ void BufferedReader::ReadAsync(size_t i) {
         auto cpu_place = cpu[i].place();
         auto cpu_ptr = cpu[i].data();
         auto mlu_ptr = mlu_ptrs[i];
-        auto size =
-            cpu[i].numel() * paddle::framework::DataTypeSize(cpu[i].dtype());
+        auto size = cpu[i].numel() * phi::SizeOf(cpu[i].dtype());
         if ((platform::is_mlu_place(cpu_place))) {
           memory::Copy(
               place_, mlu_ptr, cpu_place, cpu_ptr, size, stream_.get());
@@ -417,8 +413,7 @@ void BufferedReader::ReadAsync(size_t i) {
         auto cpu_place = cpu[i].place();
         auto cpu_ptr = cpu[i].data();
         auto xpu_ptr = xpu_ptrs[i];
-        auto size =
-            cpu[i].numel() * paddle::framework::DataTypeSize(cpu[i].dtype());
+        auto size = cpu[i].numel() * phi::SizeOf(cpu[i].dtype());
         // TODO(zhanghuan) for now hardware not support xpu_memcpy_async, maybe
         // KL3
         if ((platform::is_xpu_place(cpu_place))) {
@@ -471,8 +466,7 @@ void BufferedReader::ReadAsync(size_t i) {
         auto cpu_place = cpu[i].place();
         auto cpu_ptr = cpu[i].data();
         auto custom_device_ptr = custom_device_ptrs[i];
-        auto size =
-            cpu[i].numel() * paddle::framework::DataTypeSize(cpu[i].dtype());
+        auto size = cpu[i].numel() * phi::SizeOf(cpu[i].dtype());
         if ((platform::is_custom_place(cpu_place))) {
           memory::Copy(place_, custom_device_ptr, cpu_place, cpu_ptr, size);
           custom_device_stream_->Synchronize();
diff --git a/paddle/fluid/platform/device/gpu/cuda/cudnn_desc.h b/paddle/fluid/platform/device/gpu/cuda/cudnn_desc.h
index de8c30efe5a4e..677dc49cce4b6 100644
--- a/paddle/fluid/platform/device/gpu/cuda/cudnn_desc.h
+++ b/paddle/fluid/platform/device/gpu/cuda/cudnn_desc.h
@@ -23,9 +23,9 @@
 #include <string>
 #include <vector>
 
-#include "paddle/fluid/framework/convert_utils.h"
 #include "paddle/fluid/platform/device/gpu/cuda/cudnn_helper.h"
 #include "paddle/fluid/platform/device_context.h"
+#include "paddle/phi/core/utils/data_type.h"
 
 namespace phi {
 class DenseTensor;
@@ -37,7 +37,7 @@ namespace platform {
 template <typename T>
 inline cudnnDataType_t ToCudnnDataType(const T& t) {
   auto type = framework::ToDataType(t);
-  return ToCudnnDataType(type);
+  return ToCudnnDataType(phi::TransToPhiDataType(type));
 }
 
 template <typename T>
@@ -68,21 +68,20 @@ inline std::vector<T> TransformDimOrder(const std::vector<T>& dims) {
 }
 
 template <>
-inline cudnnDataType_t ToCudnnDataType(
-    const framework::proto::VarType::Type& t) {
+inline cudnnDataType_t ToCudnnDataType(const phi::DataType& t) {
   cudnnDataType_t type = CUDNN_DATA_FLOAT;
   switch (t) {
-    case framework::proto::VarType::FP16:
+    case phi::DataType::FLOAT16:
       type = CUDNN_DATA_HALF;
       break;
-    case framework::proto::VarType::FP32:
+    case phi::DataType::FLOAT32:
       type = CUDNN_DATA_FLOAT;
       break;
-    case framework::proto::VarType::FP64:
+    case phi::DataType::FLOAT64:
       type = CUDNN_DATA_DOUBLE;
       break;
 #if CUDNN_VERSION_MIN(8, 1, 0)
-    case framework::proto::VarType::BF16:
+    case phi::DataType::BFLOAT16:
       type = CUDNN_DATA_BFLOAT16;
       break;
 #endif
@@ -152,12 +151,12 @@ class TensorDescriptor {
     if (groups > 1) {
       dims_with_group[1] = dims_with_group[1] / groups;
     }
-    PADDLE_ENFORCE_GPU_SUCCESS(dynload::cudnnSetTensorNdDescriptor(
-        desc_.get(),
-        ToCudnnDataType(framework::TransToProtoVarType(tensor.dtype())),
-        dims_with_group.size(),
-        dims_with_group.data(),
-        strides.data()));
+    PADDLE_ENFORCE_GPU_SUCCESS(
+        dynload::cudnnSetTensorNdDescriptor(desc_.get(),
+                                            ToCudnnDataType(tensor.dtype()),
+                                            dims_with_group.size(),
+                                            dims_with_group.data(),
+                                            strides.data()));
   }
 
   void set(const std::vector<int>& dims,
@@ -179,8 +178,7 @@ class TensorDescriptor {
 
   void set(const phi::DenseTensor& tensor, const cudnnTensorFormat_t format) {
     auto dims = phi::vectorize<int>(tensor.dims());
-    auto dtype =
-        ToCudnnDataType(framework::TransToProtoVarType(tensor.dtype()));
+    auto dtype = ToCudnnDataType(tensor.dtype());
     set(dims, format, dtype);
   }
 
@@ -232,8 +230,7 @@ class FilterDescriptor {
            const cudnnTensorFormat_t format,
            const int groups = 1) {
     auto dims = phi::vectorize<int>(tensor.dims());
-    auto dtype =
-        ToCudnnDataType(framework::TransToProtoVarType(tensor.dtype()));
+    auto dtype = ToCudnnDataType(tensor.dtype());
     set(dims, format, dtype, groups);
   }
 
diff --git a/paddle/fluid/platform/device/gpu/rocm/miopen_desc.h b/paddle/fluid/platform/device/gpu/rocm/miopen_desc.h
index 8faae285e49e3..1ce4df05be64e 100644
--- a/paddle/fluid/platform/device/gpu/rocm/miopen_desc.h
+++ b/paddle/fluid/platform/device/gpu/rocm/miopen_desc.h
@@ -25,6 +25,7 @@
 
 #include "paddle/fluid/platform/device/gpu/rocm/miopen_helper.h"
 #include "paddle/fluid/platform/device_context.h"
+#include "paddle/phi/core/utils/data_type.h"
 
 namespace phi {
 class DenseTensor;
@@ -36,7 +37,7 @@ namespace platform {
 template <typename T>
 inline miopenDataType_t ToCudnnDataType(const T& t) {
   auto type = framework::ToDataType(t);
-  return ToCudnnDataType(type);
+  return ToCudnnDataType(phi::TransToPhiDataType(type));
 }
 
 inline std::vector<int> TransformDimOrder(const std::vector<int>& dims) {
@@ -63,14 +64,13 @@ inline std::vector<int> TransformDimOrder(const std::vector<int>& dims) {
 }
 
 template <>
-inline miopenDataType_t ToCudnnDataType(
-    const framework::proto::VarType::Type& t) {
+inline miopenDataType_t ToCudnnDataType(const phi::DataType& t) {
   miopenDataType_t type = miopenFloat;
   switch (t) {
-    case framework::proto::VarType::FP16:
+    case phi::DataType::FLOAT16:
       type = miopenHalf;
       break;
-    case framework::proto::VarType::FP32:
+    case phi::DataType::FLOAT32:
       type = miopenFloat;
       break;
     default:
@@ -142,7 +142,7 @@ class TensorDescriptor {
     }
     PADDLE_ENFORCE_GPU_SUCCESS(dynload::miopenSetTensorDescriptor(
         (miopenTensorDescriptor_t)(desc_.get()),
-        ToCudnnDataType(framework::TransToProtoVarType(tensor.dtype())),
+        ToCudnnDataType(tensor.dtype()),
         static_cast<int>(dims_with_group.size()),
         const_cast<int*>(dims_with_group.data()),
         const_cast<int*>(strides.data())));
@@ -166,7 +166,7 @@ class TensorDescriptor {
     }
     PADDLE_ENFORCE_GPU_SUCCESS(dynload::miopenSetTensorDescriptor(
         (miopenTensorDescriptor_t)(desc_.get()),
-        ToCudnnDataType(framework::TransToProtoVarType(tensor.dtype())),
+        ToCudnnDataType(tensor.dtype()),
         static_cast<int>(dims_with_group.size()),
         const_cast<int*>(dims_with_group.data()),
         const_cast<int*>(strides.data())));
@@ -214,7 +214,7 @@ class FilterDescriptor {
     }
     PADDLE_ENFORCE_GPU_SUCCESS(dynload::miopenSetTensorDescriptor(
         (miopenTensorDescriptor_t)(desc_.get()),
-        ToCudnnDataType(framework::TransToProtoVarType(tensor.dtype())),
+        ToCudnnDataType(tensor.dtype()),
         static_cast<int>(dims_with_group.size()),
         const_cast<int*>(dims_with_group.data()),
         const_cast<int*>(strides.data())));
diff --git a/paddle/fluid/pybind/eager_functions.cc b/paddle/fluid/pybind/eager_functions.cc
index 3389daf330c7c..14368a9d99cad 100644
--- a/paddle/fluid/pybind/eager_functions.cc
+++ b/paddle/fluid/pybind/eager_functions.cc
@@ -79,7 +79,7 @@ class EagerNumpyAllocation : public phi::Allocation {
   explicit EagerNumpyAllocation(PyObject* numpy_data, phi::DataType dtype)
       : Allocation(
             static_cast<void*>(pybind11::detail::array_proxy(numpy_data)->data),
-            framework::DataTypeSize(dtype) * PyArray_Size_(numpy_data),
+            phi::SizeOf(dtype) * PyArray_Size_(numpy_data),
             paddle::platform::CPUPlace()),
         arr_(numpy_data) {
     PADDLE_ENFORCE_NOT_NULL(
diff --git a/paddle/fluid/pybind/eager_method.cc b/paddle/fluid/pybind/eager_method.cc
index 3c52a705fc506..b2de0bf3d88e8 100644
--- a/paddle/fluid/pybind/eager_method.cc
+++ b/paddle/fluid/pybind/eager_method.cc
@@ -115,7 +115,7 @@ static PyObject* tensor_method_numpy(TensorObject* self,
   }
   auto tensor_dims = self->tensor.shape();
   auto numpy_dtype = TensorDtype2NumpyDtype(self->tensor.type());
-  auto sizeof_dtype = paddle::framework::DataTypeSize(self->tensor.type());
+  auto sizeof_dtype = phi::SizeOf(self->tensor.type());
   Py_intptr_t py_dims[paddle::framework::DDim::kMaxRank];
   Py_intptr_t py_strides[paddle::framework::DDim::kMaxRank];
   size_t numel = 1;
@@ -202,8 +202,7 @@ static PyObject* tensor_method_numpy(TensorObject* self,
       paddle::platform::GpuMemcpySync(
           pybind11::detail::array_proxy(array)->data,
           dense_tensor->data(),
-          paddle::framework::DataTypeSize(dense_tensor->dtype()) *
-              dense_tensor->numel(),
+          phi::SizeOf(dense_tensor->dtype()) * dense_tensor->numel(),
           kind);
     } else {
       VLOG(6) << "Getting DenseTensor's numpy value";
@@ -212,8 +211,7 @@ static PyObject* tensor_method_numpy(TensorObject* self,
       paddle::platform::GpuMemcpySync(
           pybind11::detail::array_proxy(array)->data,
           dense_tensor->data(),
-          paddle::framework::DataTypeSize(dense_tensor->dtype()) *
-              dense_tensor->numel(),
+          phi::SizeOf(dense_tensor->dtype()) * dense_tensor->numel(),
           kind);
     }
 #endif
@@ -257,8 +255,7 @@ static PyObject* tensor_method_numpy(TensorObject* self,
           ->MemoryCopyD2H(
               pybind11::detail::array_proxy(array)->data,
               dense_tensor->data(),
-              paddle::framework::DataTypeSize(dense_tensor->dtype()) *
-                  dense_tensor->numel());
+              phi::SizeOf(dense_tensor->dtype()) * dense_tensor->numel());
     } else {
       VLOG(6) << "Getting DenseTensor's numpy value";
       auto dense_tensor =
@@ -267,8 +264,7 @@ static PyObject* tensor_method_numpy(TensorObject* self,
           ->MemoryCopyD2H(
               pybind11::detail::array_proxy(array)->data,
               dense_tensor->data(),
-              paddle::framework::DataTypeSize(dense_tensor->dtype()) *
-                  dense_tensor->numel());
+              phi::SizeOf(dense_tensor->dtype()) * dense_tensor->numel());
     }
 #endif
   } else {
@@ -1660,7 +1656,7 @@ static PyObject* tensor_method_element_size(TensorObject* self,
                                             PyObject* args,
                                             PyObject* kwargs) {
   EAGER_TRY
-  uint32_t element_size = framework::DataTypeSize(self->tensor.dtype());
+  uint32_t element_size = phi::SizeOf(self->tensor.dtype());
 
   return ToPyObject(element_size);
   EAGER_CATCH_AND_THROW_RETURN_NULL
diff --git a/paddle/fluid/pybind/imperative.cc b/paddle/fluid/pybind/imperative.cc
index 1eb5f8bd4764c..7adec4dca2bb7 100644
--- a/paddle/fluid/pybind/imperative.cc
+++ b/paddle/fluid/pybind/imperative.cc
@@ -562,7 +562,7 @@ void BindImperative(py::module *m_ptr) {
               &t, array, platform::CPUPlace(), true);
           // 3. allocate shared memory
           void *data_ptr = t.data();
-          size_t data_size = t.numel() * framework::DataTypeSize(t.dtype());
+          size_t data_size = t.numel() * phi::SizeOf(t.dtype());
           auto shared_writer_holder =
               memory::allocation::AllocateMemoryMapWriterAllocation(data_size);
           // 4. maintain mmap fd set & backup ipc_name
@@ -602,7 +602,7 @@ void BindImperative(py::module *m_ptr) {
             &t, array, platform::CPUPlace(), true);
         // 3. allocate shared memory
         void *data_ptr = t.data();
-        size_t data_size = t.numel() * framework::DataTypeSize(t.dtype());
+        size_t data_size = t.numel() * phi::SizeOf(t.dtype());
         auto shared_writer_holder =
             memory::allocation::AllocateMemoryMapWriterAllocation(data_size);
         // 4. maintain mmap fd set & backup ipc_name
diff --git a/paddle/phi/common/data_type.h b/paddle/phi/common/data_type.h
index 339f240dae268..d66c6d89fe66e 100644
--- a/paddle/phi/common/data_type.h
+++ b/paddle/phi/common/data_type.h
@@ -258,6 +258,8 @@ inline std::string DataTypeToString(const DataType& dtype) {
 
 namespace phi {
 using DataType = paddle::experimental::DataType;
+using paddle::experimental::DataTypeToString;
+using paddle::experimental::SizeOf;
 }  // namespace phi
 
 namespace paddle {
diff --git a/paddle/phi/core/compat/CMakeLists.txt b/paddle/phi/core/compat/CMakeLists.txt
index 31b1636d5b2fd..3d76cb1112c7c 100644
--- a/paddle/phi/core/compat/CMakeLists.txt
+++ b/paddle/phi/core/compat/CMakeLists.txt
@@ -9,6 +9,10 @@ cc_library(
 
 set(convert_utils_deps data_type place op_utils phi_backends)
 
+if(WITH_MKLDNN)
+  set(convert_utils_deps ${convert_utils_deps} mkldnn)
+endif()
+
 cc_library(
   convert_utils
   SRCS convert_utils.cc
diff --git a/paddle/phi/core/utils/data_type.h b/paddle/phi/core/utils/data_type.h
index 7852d87c9a293..6879c6206564c 100644
--- a/paddle/phi/core/utils/data_type.h
+++ b/paddle/phi/core/utils/data_type.h
@@ -22,31 +22,6 @@ limitations under the License. */
 #include "paddle/phi/core/enforce.h"
 namespace phi {
 
-// Here we can't depend on the fluid proto::VarType, so we use the dtype enum
-// value directly. See also `assign_value_sig.cc`.
-// proto::VarType::INT16 -> 1  -> phi::DataType::INT16
-// proto::VarType::INT32 -> 2  -> phi::DataType::INT32
-// proto::VarType::INT64 -> 3  -> phi::DataType::INT64
-// proto::VarType::FP16 ->  4  -> phi::DataType::FLOAT16
-// proto::VarType::FP32 ->  5  -> phi::DataType::FLOAT32
-// proto::VarType::FP64 ->  6  -> phi::DataType::FLOAT64
-// proto::VarType::UINT8 -> 20 -> phi::DataType::UINT8
-static std::map<int, phi::DataType> var_type_map{{1, phi::DataType::INT16},
-                                                 {2, phi::DataType::INT32},
-                                                 {3, phi::DataType::INT64},
-                                                 {4, phi::DataType::FLOAT16},
-                                                 {5, phi::DataType::FLOAT32},
-                                                 {6, phi::DataType::FLOAT64},
-                                                 {20, phi::DataType::UINT8}};
-
-static std::map<phi::DataType, int> map_to_var_type{{phi::DataType::INT16, 1},
-                                                    {phi::DataType::INT32, 2},
-                                                    {phi::DataType::INT64, 3},
-                                                    {phi::DataType::FLOAT16, 4},
-                                                    {phi::DataType::FLOAT32, 5},
-                                                    {phi::DataType::FLOAT64, 6},
-                                                    {phi::DataType::UINT8, 20}};
-
 #define _PhiForEachDataTypeHelper_(callback, cpp_type, data_type) \
   callback(cpp_type, data_type);
 
@@ -136,39 +111,98 @@ inline DataType ToRealType(const DataType& type) {
   }
 }
 
-inline std::string DataType2String(DataType dtype) {
+// In some cases we need to use the conversion between phi::DataType and
+// fluid proto::VarType::Type, but can't depend on the proto::VarType::Type.
+// So here we defined an enum type ProtoDataType which corresponds to
+// proto::VarType::Type in fluid, but keeps only the data types we need.
+// Note: The ProtoDataType (defined here) and proto::VarType::Type (defined
+// in framework.pb.h) need to be modified simultaneously.
+enum ProtoDataType {
+  BOOL = 0,
+  INT16 = 1,
+  INT32 = 2,
+  INT64 = 3,
+  FP16 = 4,
+  FP32 = 5,
+  FP64 = 6,
+  UINT8 = 20,
+  INT8 = 21,
+  BF16 = 22,
+  COMPLEX64 = 23,
+  COMPLEX128 = 24,
+  PSTRING = 29
+};
+
+inline DataType TransToPhiDataType(const int& dtype) {
+  // Set the order of case branches according to the frequency with
+  // the data type is used
   switch (dtype) {
-    case DataType::BOOL:
-      return "bool";
+    case ProtoDataType::FP32:
+      return DataType::FLOAT32;
+    case ProtoDataType::FP64:
+      return DataType::FLOAT64;
+    case ProtoDataType::INT64:
+      return DataType::INT64;
+    case ProtoDataType::INT32:
+      return DataType::INT32;
+    case ProtoDataType::INT8:
+      return DataType::INT8;
+    case ProtoDataType::UINT8:
+      return DataType::UINT8;
+    case ProtoDataType::INT16:
+      return DataType::INT16;
+    case ProtoDataType::COMPLEX64:
+      return DataType::COMPLEX64;
+    case ProtoDataType::COMPLEX128:
+      return DataType::COMPLEX128;
+    case ProtoDataType::FP16:
+      return DataType::FLOAT16;
+    case ProtoDataType::BF16:
+      return DataType::BFLOAT16;
+    case ProtoDataType::BOOL:
+      return DataType::BOOL;
+    case ProtoDataType::PSTRING:
+      return DataType::PSTRING;
+    default:
+      return DataType::UNDEFINED;
+  }
+}
+
+inline int TransToProtoVarType(const DataType& dtype) {
+  // Set the order of case branches according to the frequency with
+  // the data type is used
+  switch (dtype) {
+    case DataType::FLOAT32:
+      return ProtoDataType::FP32;
+    case DataType::FLOAT64:
+      return ProtoDataType::FP64;
+    case DataType::INT64:
+      return ProtoDataType::INT64;
+    case DataType::INT32:
+      return ProtoDataType::INT32;
     case DataType::INT8:
-      return "int8";
+      return ProtoDataType::INT8;
     case DataType::UINT8:
-      return "uint8";
+      return ProtoDataType::UINT8;
     case DataType::INT16:
-      return "int16";
-    case DataType::INT32:
-      return "int32";
-    case DataType::INT64:
-      return "int64";
-    case DataType::FLOAT16:
-      return "float16";
-    case DataType::FLOAT32:
-      return "float32";
-    case DataType::FLOAT64:
-      return "float64";
+      return ProtoDataType::INT16;
     case DataType::COMPLEX64:
-      return "complex64";
+      return ProtoDataType::COMPLEX64;
     case DataType::COMPLEX128:
-      return "complex128";
-    case DataType::PSTRING:
-      return "pstring";
+      return ProtoDataType::COMPLEX128;
+    case DataType::FLOAT16:
+      return ProtoDataType::FP16;
     case DataType::BFLOAT16:
-      return "bfloat16";
+      return ProtoDataType::BF16;
+    case DataType::BOOL:
+      return ProtoDataType::BOOL;
+    case DataType::PSTRING:
+      return ProtoDataType::PSTRING;
     default:
-      PADDLE_THROW(
-          errors::InvalidArgument("Unknow phi::DataType, the int value = %d.",
-                                  static_cast<int>(dtype)));
-      return "";
+      PADDLE_THROW(phi::errors::Unimplemented(
+          "Unsupported data type `%s` when casting it into "
+          "paddle data type.",
+          dtype));
   }
 }
 
diff --git a/paddle/phi/infermeta/unary.cc b/paddle/phi/infermeta/unary.cc
index 5f602a134ecd9..f2c3873d81e5c 100644
--- a/paddle/phi/infermeta/unary.cc
+++ b/paddle/phi/infermeta/unary.cc
@@ -133,9 +133,9 @@ void ArgMinMaxInferMeta(const MetaTensor& x,
       phi::errors::InvalidArgument(
           "The attribute of dtype in argmin/argmax must be [%s] or [%s], but "
           "received [%s]",
-          phi::DataType2String(DataType::INT32),
-          phi::DataType2String(DataType::INT64),
-          phi::DataType2String(var_type_map[dtype])));
+          phi::DataTypeToString(DataType::INT32),
+          phi::DataTypeToString(DataType::INT64),
+          phi::DataTypeToString(phi::TransToPhiDataType(dtype))));
 
   if (!config.is_runtime && axis.FromTensor()) {
     std::vector<int64_t> vec;
@@ -177,7 +177,7 @@ void ArgMinMaxInferMeta(const MetaTensor& x,
   auto x_rank = x_dims.size();
   if (int_axis < 0) int_axis += x_rank;
   if (config.is_runtime) {
-    if (dtype == map_to_var_type[DataType::INT32]) {
+    if (dtype == phi::TransToProtoVarType(DataType::INT32)) {
       int64_t all_element_num = 0;
       if (flatten) {
         all_element_num = phi::product(x_dims);
diff --git a/paddle/phi/kernels/cpu/arg_min_max_kernel.cc b/paddle/phi/kernels/cpu/arg_min_max_kernel.cc
index 999cb16620d26..5c7e79c9e8ade 100644
--- a/paddle/phi/kernels/cpu/arg_min_max_kernel.cc
+++ b/paddle/phi/kernels/cpu/arg_min_max_kernel.cc
@@ -149,7 +149,7 @@ void ArgMinMaxKernel(const Context& dev_ctx,
     return;
   }
   phi::VisitDataTypeTiny(
-      var_type_map[dtype],
+      phi::TransToPhiDataType(dtype),
       VisitDataArgMinMaxFunctor<Context, T, EnumArgMinMaxValue>(
           dev_ctx, x, axis.to<int64_t>(), keepdims, flatten, out));
 }
diff --git a/paddle/phi/kernels/cpu/index_sample_grad_kernel.cc b/paddle/phi/kernels/cpu/index_sample_grad_kernel.cc
index 50f2c3267fbc5..42aef3cc2482b 100644
--- a/paddle/phi/kernels/cpu/index_sample_grad_kernel.cc
+++ b/paddle/phi/kernels/cpu/index_sample_grad_kernel.cc
@@ -81,9 +81,9 @@ void IndexSampleGradKernel(const Context& ctx,
                     errors::InvalidArgument(
                         "Input(Index) holds the wrong type, it holds %s, but "
                         "desires to be %s or %s",
-                        phi::DataType2String(index_type),
-                        phi::DataType2String(DataType::INT32),
-                        phi::DataType2String(DataType::INT64)));
+                        phi::DataTypeToString(index_type),
+                        phi::DataTypeToString(DataType::INT32),
+                        phi::DataTypeToString(DataType::INT64)));
   if (index_type == DataType::INT32) {
     IndexSampleGradInner<T, Context, int>(ctx, out_grad, index, x_grad);
   } else if (index_type == DataType::INT64) {
diff --git a/paddle/phi/kernels/cpu/index_sample_kernel.cc b/paddle/phi/kernels/cpu/index_sample_kernel.cc
index 11e24b10b153b..e51d06c442408 100644
--- a/paddle/phi/kernels/cpu/index_sample_kernel.cc
+++ b/paddle/phi/kernels/cpu/index_sample_kernel.cc
@@ -94,9 +94,9 @@ void IndexSampleKernel(const Context &ctx,
                     errors::InvalidArgument(
                         "Input(Index) holds the wrong type, it holds %s, but "
                         "desires to be %s or %s",
-                        phi::DataType2String(index_type),
-                        phi::DataType2String(DataType::INT32),
-                        phi::DataType2String(DataType::INT64)));
+                        phi::DataTypeToString(index_type),
+                        phi::DataTypeToString(DataType::INT32),
+                        phi::DataTypeToString(DataType::INT64)));
   if (index_type == DataType::INT32) {
     IndexSampleInner<T, Context, int>(ctx, x, index, out);
   } else if (index_type == DataType::INT64) {
diff --git a/paddle/phi/kernels/cpu/unique_consecutive_kernel.cc b/paddle/phi/kernels/cpu/unique_consecutive_kernel.cc
index 07df5f1f566a4..2e1dd3e4ecbf1 100644
--- a/paddle/phi/kernels/cpu/unique_consecutive_kernel.cc
+++ b/paddle/phi/kernels/cpu/unique_consecutive_kernel.cc
@@ -32,7 +32,7 @@ void UniqueConsecutiveKernel(const Context& dev_ctx,
                              DenseTensor* out,
                              DenseTensor* index,
                              DenseTensor* counts) {
-  auto data_type = var_type_map[dtype];
+  auto data_type = phi::TransToPhiDataType(dtype);
   if (data_type == phi::DataType::INT32) {
     PADDLE_ENFORCE_LE(
         x.numel(),
diff --git a/paddle/phi/kernels/funcs/unique_functor.h b/paddle/phi/kernels/funcs/unique_functor.h
index 2c713243904eb..edd3935ef7603 100644
--- a/paddle/phi/kernels/funcs/unique_functor.h
+++ b/paddle/phi/kernels/funcs/unique_functor.h
@@ -82,9 +82,9 @@ struct UniqueOpFunctor {
                         phi::errors::InvalidArgument(
                             "Index holds the wrong type, it holds %s, "
                             "but desires to be %s or %s",
-                            phi::DataType2String(index_type),
-                            phi::DataType2String(DataType::INT32),
-                            phi::DataType2String(DataType::INT64)));
+                            phi::DataTypeToString(index_type),
+                            phi::DataTypeToString(DataType::INT32),
+                            phi::DataTypeToString(DataType::INT64)));
 
       if (index_type == DataType::INT32) {
         for (auto i = 0; i < in_->numel(); ++i) {
diff --git a/paddle/phi/kernels/gpu/arg_min_max_kernel.cu b/paddle/phi/kernels/gpu/arg_min_max_kernel.cu
index 0cbf206bb3689..5764ba54b397a 100644
--- a/paddle/phi/kernels/gpu/arg_min_max_kernel.cu
+++ b/paddle/phi/kernels/gpu/arg_min_max_kernel.cu
@@ -215,7 +215,7 @@ void ArgMinMaxOpCUDAKernel(const Context& dev_ctx,
     return;
   }
   phi::VisitDataTypeTiny(
-      var_type_map[dtype],
+      phi::TransToPhiDataType(dtype),
       VisitDataCudaArgMinMaxFunctor<Context, T, Reducer>(
           dev_ctx, x, axis.to<int64_t>(), keepdims, flatten, out));
 }
diff --git a/paddle/phi/kernels/gpu/index_sample_grad_kernel.cu b/paddle/phi/kernels/gpu/index_sample_grad_kernel.cu
index 5193d0a7ab05c..b016cf20b1332 100644
--- a/paddle/phi/kernels/gpu/index_sample_grad_kernel.cu
+++ b/paddle/phi/kernels/gpu/index_sample_grad_kernel.cu
@@ -75,9 +75,9 @@ void IndexSampleGradKernel(const Context& ctx,
                     errors::InvalidArgument(
                         "Input(Index) holds the wrong type, it holds %s, but "
                         "desires to be %s or %s",
-                        phi::DataType2String(index_type),
-                        phi::DataType2String(DataType::INT32),
-                        phi::DataType2String(DataType::INT64)));
+                        phi::DataTypeToString(index_type),
+                        phi::DataTypeToString(DataType::INT32),
+                        phi::DataTypeToString(DataType::INT64)));
 
   auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
   auto input_num = x.numel();
diff --git a/paddle/phi/kernels/gpu/index_sample_kernel.cu b/paddle/phi/kernels/gpu/index_sample_kernel.cu
index dd1b4aa97d2b6..31fe1ff2a02f2 100644
--- a/paddle/phi/kernels/gpu/index_sample_kernel.cu
+++ b/paddle/phi/kernels/gpu/index_sample_kernel.cu
@@ -64,9 +64,9 @@ void IndexSampleKernel(const Context& ctx,
                     errors::InvalidArgument(
                         "Input(Index) holds the wrong type, it holds %s, but "
                         "desires to be %s or %s",
-                        phi::DataType2String(index_type),
-                        phi::DataType2String(DataType::INT32),
-                        phi::DataType2String(DataType::INT64)));
+                        phi::DataTypeToString(index_type),
+                        phi::DataTypeToString(DataType::INT32),
+                        phi::DataTypeToString(DataType::INT64)));
   const T* in_data = x.data<T>();
   T* out_data = ctx.template Alloc<T>(out);
   auto stream = reinterpret_cast<const phi::GPUContext&>(ctx).stream();
diff --git a/paddle/phi/kernels/gpu/unique_consecutive_kernel.cu b/paddle/phi/kernels/gpu/unique_consecutive_kernel.cu
index 9eb9309bb496f..9c04ab511d04f 100644
--- a/paddle/phi/kernels/gpu/unique_consecutive_kernel.cu
+++ b/paddle/phi/kernels/gpu/unique_consecutive_kernel.cu
@@ -33,7 +33,7 @@ void UniqueConsecutiveKernel(const Context& dev_ctx,
                              DenseTensor* out,
                              DenseTensor* index,
                              DenseTensor* counts) {
-  auto data_type = var_type_map[dtype];
+  auto data_type = phi::TransToPhiDataType(dtype);
   if (data_type == phi::DataType::INT32) {
     PADDLE_ENFORCE_LE(
         x.numel() + 1,
diff --git a/paddle/phi/kernels/gpudnn/conv_cudnn_frontend.h b/paddle/phi/kernels/gpudnn/conv_cudnn_frontend.h
index e45ff63d41b3a..368b5585b291e 100644
--- a/paddle/phi/kernels/gpudnn/conv_cudnn_frontend.h
+++ b/paddle/phi/kernels/gpudnn/conv_cudnn_frontend.h
@@ -17,11 +17,11 @@ limitations under the License. */
 
 #include <vector>
 
-#include "paddle/fluid/framework/convert_utils.h"
 #include "paddle/fluid/platform/device/gpu/cuda/cudnn_desc.h"
 #include "paddle/phi/backends/dynload/cudnn_frontend.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
+#include "paddle/phi/core/utils/data_type.h"
 #include "paddle/phi/kernels/autotune/cache.h"
 #include "paddle/phi/kernels/autotune/switch_autotune.h"
 
@@ -95,8 +95,7 @@ class CudnnFrontendConvHelper {
         .setStrides(strides.size(), strides.data())
         .setId(id)
         .setAlignment(GetAlignment(tensor))
-        .setDataType(paddle::platform::ToCudnnDataType(
-            paddle::framework::TransToProtoVarType(tensor->dtype())))
+        .setDataType(paddle::platform::ToCudnnDataType(tensor->dtype()))
         .build();
   }