!45801 Supports the ability of the print operator on the CPU to print tuple constants.

Merge pull request !45801 from Margaret_wangrui/print_cpu_tuple

.jenkins/check/config/whitelizard.txt

@@ -11,6 +11,7 @@ mindspore/mindspore/lite/providers/nnie_proposal/src/proposal.cc:mindspore::prop
 mindspore/mindspore/core/abstract/ops/primitive_infer_map.cc:mindspore::abstract::GetPrimitiveToEvalImplMap
 mindspore/mindspore/core/abstract/ops/primitive_infer_map.cc:mindspore::abstract::GetHostDependsMap
 mindspore/mindspore/core/ir/tensor.cc:mindspore::tensor::MakeTensorData
+mindspore/mindspore/ccsrc/kernel/common_utils.cc:mindspore::kernel::UnitSizeInBytes
 mindspore/mindspore/ccsrc/frontend/optimizer/irpass.cc:mindspore::opt::irpass::OptimizeIRPassLib::OptimizeIRPassLib
 mindspore/mindspore/ccsrc/frontend/parallel/ops_info/gather_v2_p_info.cc:mindspore::parallel::GatherV2PInfo::CheckStrategy
 mindspore/mindspore/ccsrc/plugin/device/gpu/hal/device/gpu_kernel_runtime.cc:mindspore::device::gpu::GPUKernelRuntime::LaunchKernelDynamic

mindspore/ccsrc/kernel/common_utils.cc

@@ -1113,6 +1113,9 @@ size_t UnitSizeInBytes(const mindspore::TypeId &t) {
     case kNumberTypeComplex128:
       bytes = sizeof(double) * complex_factor;
       break;
+    case kObjectTypeString:
+      bytes = sizeof(std::string);
+      break;
     case kNumberTypeInt4:
     default:
       MS_LOG(EXCEPTION) << "Invalid types for UnitSizeInBytes : " << TypeIdToString(t);

mindspore/ccsrc/plugin/device/cpu/hal/hardware/cpu_device_context.cc

@@ -20,6 +20,7 @@
 #include "plugin/device/cpu/hal/device/cpu_device_address.h"
 #include "plugin/device/cpu/hal/device/cpu_memory_manager.h"
 #include "plugin/device/cpu/optimizer/reg_cpu_const_input_to_attr.h"
+#include "plugin/device/cpu/optimizer/print_value_type.h"
 #include "plugin/device/cpu/hal/hardware/cpu_somas.h"
 #ifdef ENABLE_AKG
 #include "plugin/device/cpu/kernel/akg/akg_cpu_kernel_build.h"
@@ -195,6 +196,7 @@ void CPUKernelExecutor::OptimizeGraphImpl(const KernelGraphPtr &graph) const {
   pm->AddPass(std::make_shared<opt::EraseVisitAttr>());
   pm->AddPass(std::make_shared<opt::InsertTensorMoveForCommunication>());
   pm->AddPass(std::make_shared<opt::AddTrainingAttr>());
+  pm->AddPass(std::make_shared<opt::PrintValueType>("print_value_type"));
   optimizer->AddPassManager(pm);
   (void)optimizer->Optimize(graph);
   graph->SetExecOrderByDefault();

mindspore/ccsrc/plugin/device/cpu/hal/hardware/cpu_session.cc

@@ -27,6 +27,7 @@
 #include "plugin/factory/ms_factory.h"
 #include "runtime/device/kernel_runtime.h"
 #include "plugin/device/cpu/kernel/cpu_kernel.h"
+#include "plugin/device/cpu/optimizer/print_value_type.h"
 #ifdef ENABLE_AKG
 #include "plugin/device/cpu/kernel/akg/akg_cpu_kernel_build.h"
 #endif
@@ -105,6 +106,7 @@ void CPUSession::Optimize(const std::shared_ptr<KernelGraph> &kernel_graph) {
   pm->AddPass(std::make_shared<opt::InsertFormatTransformOpCPU>("insert_format_transform_op_cpu"));
   pm->AddPass(std::make_shared<opt::InsertCastCPU>("insert_cast"));
   pm->AddPass(std::make_shared<opt::EraseVisitAttr>());
+  pm->AddPass(std::make_shared<opt::PrintValueType>("print_value_type"));
   optimizer->AddPassManager(pm);
   (void)optimizer->Optimize(kernel_graph);
   kernel_graph->SetExecOrderByDefault();

mindspore/ccsrc/plugin/device/cpu/kernel/print_cpu_kernel.cc

@@ -15,11 +15,13 @@
  */
 
 #include "plugin/device/cpu/kernel/print_cpu_kernel.h"
+#include <functional>
 #include <algorithm>
 #include <utility>
 #include <string>
 #include <complex>
 #include "ir/tensor.h"
+#include "ops/print.h"
 #include "plugin/device/cpu/hal/device/cpu_device_address.h"
 
 namespace mindspore {
@@ -36,6 +38,14 @@ bool PrintCpuKernelMod::Init(const BaseOperatorPtr &base_operator, const std::ve
     TypeId type = inputs[i]->GetDtype();
     (void)data_types_.emplace_back(type);
   }
+  auto kernel_ptr = std::dynamic_pointer_cast<ops::Print>(base_operator);
+  if (kernel_ptr->HasAttr("value_type")) {
+    auto value_type = kernel_ptr->get_value_type();
+    auto value_type_pos = kernel_ptr->get_value_type_pos();
+    for (size_t i = 0; i < value_type.size(); i++) {
+      value_type_[value_type_pos[i]] = value_type[i];
+    }
+  }
   return true;
 }
 
@@ -46,9 +56,9 @@ int PrintCpuKernelMod::Resize(const BaseOperatorPtr &base_operator, const std::v
   if (ret != KRET_OK) {
     return ret;
   }
-
   input_sizes_.clear();
   input_shapes_.clear();
+  input_info_.clear();
   for (size_t i = 0; i < inputs.size(); ++i) {
     MS_EXCEPTION_IF_NULL(inputs[i]);
     auto input_shape = inputs[i]->GetShapeVector();
@@ -57,7 +67,11 @@ int PrintCpuKernelMod::Resize(const BaseOperatorPtr &base_operator, const std::v
     for (size_t j = 0; j < input_shape.size(); ++j) {
       size *= input_shape[j];
     }
+    auto type_id = inputs[i]->GetDtype();
+    size_t unit_size = UnitSizeInBytes(type_id);
+    auto size_in_byte = std::accumulate(input_shape.begin(), input_shape.end(), unit_size, std::multiplies<size_t>());
     (void)input_sizes_.emplace_back(LongToSize(size));
+    input_info_.push_back(std::make_tuple(size_in_byte, type_id));
   }
   return ret;
 }
@@ -91,8 +105,20 @@ void PrintCpuKernelMod::LaunchKernel(size_t index, const std::vector<kernel::Add
       std::cout << *num << std::endl;
     }
   } else {
+    TypeId type_id = std::get<1>(input_info_[index]);
     Tensor tensor(data_types_[index], input_shapes_[index], inputs[index]->addr, input_sizes_[index] * sizeof(T));
-    std::cout << tensor.ToStringNoLimit() << std::endl;
+    if (value_type_.count(index) > 0) {
+      // not a tensor
+      auto out = tensor.data().ToString(type_id, input_shapes_[index], true);
+      if (value_type_[index] != 0) {
+        // tuple, not scalar
+        (void)std::replace(out.begin(), out.end(), '[', '(');
+        (void)std::replace(out.begin(), out.end(), ']', ')');
+      }
+      std::cout << out << std::endl;
+    } else {
+      std::cout << tensor.ToStringNoLimit() << std::endl;
+    }
   }
 }
 

mindspore/ccsrc/plugin/device/cpu/kernel/print_cpu_kernel.h

@@ -21,6 +21,8 @@
 #include <vector>
 #include <utility>
 #include <map>
+#include <tuple>
+#include <unordered_map>
 #include "plugin/device/cpu/kernel/cpu_kernel.h"
 #include "plugin/factory/ms_factory.h"
 
@@ -54,6 +56,9 @@ class PrintCpuKernelMod : public NativeCpuKernelMod {
   std::vector<ShapeVector> input_shapes_;
   std::vector<size_t> input_sizes_;
   std::vector<TypeId> data_types_;
+
+  std::unordered_map<int64_t, int64_t> value_type_;
+  std::vector<std::tuple<size_t, TypeId>> input_info_;
 };
 }  // namespace kernel
 }  // namespace mindspore

mindspore/ccsrc/plugin/device/cpu/optimizer/print_value_type.cc

@@ -0,0 +1,144 @@
+/**
+ * Copyright 2022 Huawei Technologies Co., Ltd
+ *
+ * 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 "plugin/device/cpu/optimizer/print_value_type.h"
+
+#include <memory>
+#include <vector>
+#include <string>
+#include <utility>
+#include <algorithm>
+#include "backend/common/session/anf_runtime_algorithm.h"
+#include "include/common/utils/anfalgo.h"
+#include "ir/primitive.h"
+#include "include/common/utils/utils.h"
+#include "backend/common/optimizer/helper.h"
+
+namespace mindspore {
+namespace opt {
+kernel::KernelBuildInfoPtr GenerateKernelBuildInfo(CNodePtr node) {
+  std::vector<std::string> inputs_format;
+  std::vector<std::string> outputs_format;
+  std::vector<TypeId> inputs_type;
+  std::vector<TypeId> outputs_type;
+  kernel::KernelBuildInfo::KernelBuildInfoBuilder builder;
+
+  size_t input_num = common::AnfAlgo::GetInputTensorNum(node);
+  for (size_t input_index = 0; input_index < input_num; input_index++) {
+    inputs_format.push_back(kOpFormat_DEFAULT);
+    inputs_type.push_back(common::AnfAlgo::GetPrevNodeOutputInferDataType(node, input_index));
+  }
+  size_t output_num = common::AnfAlgo::GetOutputTensorNum(node);
+  for (size_t output_index = 0; output_index < output_num; output_index++) {
+    outputs_format.push_back(kOpFormat_DEFAULT);
+    outputs_type.push_back(common::AnfAlgo::GetOutputInferDataType(node, output_index));
+  }
+
+  builder.SetInputsFormat(inputs_format);
+  builder.SetOutputsFormat(outputs_format);
+  builder.SetInputsDeviceType(inputs_type);
+  builder.SetOutputsDeviceType(outputs_type);
+  return builder.Build();
+}
+
+bool GetOptList(const std::vector<AnfNodePtr> &node_list, std::vector<AnfNodePtr> *opt_list,
+                std::vector<std::vector<std::pair<int64_t, int64_t>>> *not_tensor_pos_vec) {
+  MS_EXCEPTION_IF_NULL(opt_list);
+
+  for (auto &node : node_list) {
+    // {prim::kPrimPrint} reduction only applies on print with string, tensor(scalar or tuple)
+    MS_EXCEPTION_IF_NULL(node);
+    std::vector<std::pair<int64_t, int64_t>> value_type;
+    if (!IsPrimitiveCNode(node, prim::kPrimPrint)) {
+      continue;
+    }
+    size_t input_num = common::AnfAlgo::GetInputTensorNum(node);
+    for (size_t i = 0; i < input_num; i++) {
+      auto current_node = common::AnfAlgo::GetInputNode(utils::cast<CNodePtr>(node), i);
+      // not tensor(tuple, scalar, string)
+      if (current_node->cast<ValueNodePtr>() == nullptr) {
+        continue;
+      }
+
+      auto value_node = current_node->cast<ValueNodePtr>();
+      MS_EXCEPTION_IF_NULL(value_node);
+      auto shape = value_node->abstract();
+      MS_EXCEPTION_IF_NULL(shape);
+      auto shape_node = dyn_cast<abstract::Shape>(shape->GetShapeTrack());
+      if (shape_node != nullptr) {
+        // a scalar or tuple
+        auto shape_size = shape_node->shape().size();
+        if (shape_size != 0) {
+          value_type.push_back(std::make_pair(i, 1));
+        } else {
+          value_type.push_back(std::make_pair(i, 0));
+        }
+      }
+    }
+    if (value_type.size() != 0) {
+      opt_list->push_back(node);
+      not_tensor_pos_vec->push_back(value_type);
+    }
+  }
+  if (opt_list->size() == 0) {
+    return false;
+  }
+  return true;
+}
+
+bool PrintValueType::Run(const FuncGraphPtr &graph) {
+  MS_EXCEPTION_IF_NULL(graph);
+  auto manager = graph->manager();
+  MS_EXCEPTION_IF_NULL(manager);
+  std::vector<AnfNodePtr> node_list = TopoSort(graph->get_return());
+  std::vector<AnfNodePtr> opt_list;
+  // first is pos, second is type: 0 is Scalar, 1 is ValueTuple
+  std::vector<std::vector<std::pair<int64_t, int64_t>>> not_tensor_pos_vec;
+  if (!GetOptList(node_list, &opt_list, &not_tensor_pos_vec)) {
+    return false;
+  }
+  for (size_t idx = 0; idx < opt_list.size(); idx++) {
+    auto node = opt_list[idx];
+    CNodePtr cnode = utils::cast<CNodePtr>(node);
+    MS_EXCEPTION_IF_NULL(cnode);
+    auto value_type_vec = not_tensor_pos_vec[idx];
+    // split value type and pos
+    std::vector<int64_t> value_type_pos;
+    std::vector<int64_t> value_type;
+    (void)std::transform(value_type_vec.begin(), value_type_vec.end(), std::back_inserter(value_type_pos),
+                         [](const std::pair<int64_t, int64_t> &value) { return value.first; });
+    (void)std::transform(value_type_vec.begin(), value_type_vec.end(), std::back_inserter(value_type),
+                         [](const std::pair<int64_t, int64_t> &value) { return value.second; });
+
+    // hand over the attrs to new print
+    common::AnfAlgo::SetNodeAttr("value_type", MakeValue<std::vector<int64_t>>(value_type), cnode);
+    common::AnfAlgo::SetNodeAttr("value_type_pos", MakeValue<std::vector<int64_t>>(value_type_pos), cnode);
+    // set output type and shape
+    std::vector<TypeId> types;
+    std::vector<BaseShapePtr> shapes;
+    size_t output_num = common::AnfAlgo::GetOutputTensorNum(cnode);
+    for (size_t i = 0; i < output_num; i++) {
+      types.push_back(common::AnfAlgo::GetOutputInferDataType(cnode, i));
+      shapes.push_back(common::AnfAlgo::GetOutputDetailShape(cnode, i));
+    }
+    common::AnfAlgo::SetOutputTypeAndDetailShape(types, shapes, cnode.get());
+    // add build info
+    auto build_info = GenerateKernelBuildInfo(cnode);
+    AnfAlgo::SetSelectKernelBuildInfo(build_info, cnode.get());
+  }
+  return true;
+}
+}  // namespace opt
+}  // namespace mindspore

mindspore/ccsrc/plugin/device/cpu/optimizer/print_value_type.h

@@ -0,0 +1,32 @@
+/**
+ * Copyright 2022 Huawei Technologies Co., Ltd
+ *
+ * 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.
+ */
+#ifndef MINDSPORE_CCSRC_PLUGIN_DEVICE_CPU_OPTIMIZER_PRINT_VALUE_TYPE_H_
+#define MINDSPORE_CCSRC_PLUGIN_DEVICE_CPU_OPTIMIZER_PRINT_VALUE_TYPE_H_
+
+#include <string>
+#include "backend/common/optimizer/optimizer.h"
+
+namespace mindspore {
+namespace opt {
+class PrintValueType : public Pass {
+ public:
+  explicit PrintValueType(const std::string &name) : Pass("print_value_type") {}
+  ~PrintValueType() override = default;
+  bool Run(const FuncGraphPtr &graph) override;
+};
+}  // namespace opt
+}  // namespace mindspore
+#endif  // MINDSPORE_CCSRC_PLUGIN_DEVICE_CPU_OPTIMIZER_PRINT_VALUE_TYPE_H_

tests/st/ops/cpu/test_print_op.py

@@ -91,3 +91,24 @@ def test_print_op_functional(mode):
     net = PrintFunc()
     x = Tensor(np.random.randn(3, 4, 5).astype(np.float32))
     net(x)
+
+
+@pytest.mark.level0
+@pytest.mark.env_onecard
+@pytest.mark.platform_x86_cpu
+def test_print_op_tuple():
+    """
+    Feature: cpu Print op.
+    Description: test Print with tuple input.
+    Expectation: success.
+    """
+    class PrintTupleNet(nn.Cell):
+        def construct(self, x):
+            tuple_x = tuple((1, 2, 3, 4, 5))
+            ops.print_("tuple_x:", tuple_x, x, "print success!")
+            return x
+
+    context.set_context(mode=context.GRAPH_MODE, device_target="CPU")
+    net = PrintTupleNet()
+    x = Tensor([6, 7, 8, 9, 10])
+    net(x)

tests/st/ops/gpu/test_print_op.py

@@ -177,3 +177,25 @@ def test_print_dynamic_shape(mode):
     x_dyn = Tensor(shape=[None, None, None], dtype=ms.float32)
     net.set_inputs(x_dyn)
     net(x)
+
+
+@security_off_wrap
+@pytest.mark.level1
+@pytest.mark.env_onecard
+@pytest.mark.platform_x86_gpu_training
+def test_print_op_tuple():
+    """
+    Feature: cpu Print op.
+    Description: test Print with tuple input.
+    Expectation: success.
+    """
+    class PrintTupleNet(nn.Cell):
+        def construct(self, x):
+            tuple_x = tuple((1, 2, 3, 4, 5))
+            print("tuple_x:", tuple_x, x, "print success!")
+            return x
+
+    context.set_context(mode=context.GRAPH_MODE, device_target="GPU")
+    net = PrintTupleNet()
+    x = Tensor([6, 7, 8, 9, 10])
+    net(x)