!10794 Synchronize latest Ascend software suite 29 Dec 2020

From: @nicholas_yhr
Reviewed-by: @xsmq,@liucunwei
Signed-off-by: @liucunwei

graphengine

@@ -1 +1 @@
-Subproject commit c762dd5dcc207987d5b5d4ee520da3939222ec88
+Subproject commit 1b4f85776269f567d11153807ae7badc91803083

mindspore/ccsrc/backend/optimizer/ascend/ir_fission/dynamic_rnn_grad_fission_v2.cc

@@ -16,8 +16,10 @@
 #include "backend/optimizer/ascend/ir_fission/dynamic_rnn_grad_fission_v2.h"
 #include <vector>
 #include <memory>
+#include "backend/session/kernel_graph.h"
 #include "backend/session/anf_runtime_algorithm.h"
 #include "utils/trace_base.h"
+#include "utils/tensor_construct_utils.h"
 
 namespace mindspore {
 namespace opt {
@@ -46,7 +48,7 @@ void CreateTLoopNode(const FuncGraphPtr &func_graph, const CNodePtr &dynamic_rnn
 
     std::vector<size_t> output0_dims{origin_input9_shape[0], 4 * (((origin_input9_shape[1] + 15) / 16) * 16)};
     std::vector<size_t> output1_dims{input_i_shape[1], input_i_shape[2]};
-    AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat32, kNumberTypeFloat32}, {output0_dims, output1_dims},
+    AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat16, kNumberTypeFloat32}, {output0_dims, output1_dims},
                                         basic_lstm_cell_c_state_grad.get());
     AnfAlgo::SetNodeAttr("forget_bias", MakeValue(1.0f), basic_lstm_cell_c_state_grad);
     AnfAlgo::SetNodeAttr("activation", MakeValue("Tanh"), basic_lstm_cell_c_state_grad);
@@ -260,7 +262,7 @@ AnfNodePtr AddLSTMInputGradNode(const FuncGraphPtr &func_graph, const CNodePtr &
   // Create lstm_gage_concat
   auto lstm_gage_concat = func_graph->NewCNode(lstm_gage_concat_input);
   auto origin_input7_shape = AnfAlgo::GetOutputInferShape(origin_input7, 0);
-  AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat32},
+  AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat16},
                                       {{origin_input7_shape[0], origin_input7_shape[1], 4 * origin_input7_shape[2]}},
                                       lstm_gage_concat.get());
   AnfAlgo::SetNodeAttr(kAttrN, MakeValue(SizeToLong(num_split_x)), lstm_gage_concat);
@@ -413,6 +415,24 @@ AnfNodePtr CreateBatchMatMul(const FuncGraphPtr &func_graph, const AnfNodePtr &l
   return batch_matmul;
 }
 
+AnfNodePtr CreateBatchMatMul2(const FuncGraphPtr &func_graph, const AnfNodePtr &lstm_input_grad,
+                              const AnfNodePtr &node) {
+  MS_EXCEPTION_IF_NULL(func_graph);
+  // Create node
+  std::vector<AnfNodePtr> matmul_inputs = {NewValueNode(std::make_shared<Primitive>(prim::kPrimBatchMatMul->name())),
+                                           node, lstm_input_grad};
+  auto batch_matmul = func_graph->NewCNode(matmul_inputs);
+  // Set infer data type and shape
+  auto out_shape = {AnfAlgo::GetOutputInferShape(lstm_input_grad, 0)[0], IntToSize(1),
+                    AnfAlgo::GetOutputInferShape(lstm_input_grad, 0)[2]};
+  AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat16}, {out_shape}, batch_matmul.get());
+  // Set attr
+  AnfAlgo::SetNodeAttr("is_backend_insert", MakeValue(true), batch_matmul);
+  AnfAlgo::SetNodeAttr("transpose_x1", MakeValue(false), batch_matmul);
+  AnfAlgo::SetNodeAttr("transpose_x2", MakeValue(false), batch_matmul);
+  return batch_matmul;
+}
+
 AnfNodePtr CreateDwReduceSum(const FuncGraphPtr &func_graph, const CNodePtr &dynamic_rnn_grad_cnode,
                              const AnfNodePtr &batch_matmul) {
   MS_EXCEPTION_IF_NULL(func_graph);
@@ -430,18 +450,38 @@ AnfNodePtr CreateDwReduceSum(const FuncGraphPtr &func_graph, const CNodePtr &dyn
   return reduce_sum;
 }
 
+AnfNodePtr CreateValueNode(const FuncGraphPtr &func_graph, const CNodePtr &dynamic_rnn_grad_cnode) {
+  auto origin_input7 = dynamic_rnn_grad_cnode->input(8);
+  auto origin_input7_shape = AnfAlgo::GetOutputInferShape(origin_input7, 0);
+  auto t_size = origin_input7_shape[0];
+  auto n_size = origin_input7_shape[1];
+
+  std::vector<size_t> shape = {t_size, IntToSize(1), n_size};
+  std::vector<int64_t> output_shape = {SizeToLong(t_size), SizeToLong(1), SizeToLong(n_size)};
+  std::vector<int64_t> output_tensor = {(SizeToLong(n_size) + SizeToLong(15)) / SizeToLong(16) * SizeToLong(16) *
+                                        SizeToLong(16) * SizeToLong(t_size)};
+  auto tensor = TensorConstructUtils::CreateOnesTensor(kNumberTypeFloat32, output_tensor);
+  auto x_abstract = std::make_shared<abstract::AbstractTensor>(kFloat32, output_shape);
+  auto kernel_graph = func_graph->cast<KernelGraphPtr>();
+  auto value_node = kernel_graph->NewValueNode(x_abstract, tensor);
+  kernel_graph->AddValueNodeToGraph(value_node);
+  AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat32}, {shape}, value_node.get());
+  return value_node;
+}
+
 AnfNodePtr CreateDbReduceSum(const FuncGraphPtr &func_graph, const CNodePtr &dynamic_rnn_grad_cnode,
-                             const AnfNodePtr &lstm_input_grad) {
+                             const AnfNodePtr &lstm_input_grad, const AnfNodePtr &value_node) {
   MS_EXCEPTION_IF_NULL(func_graph);
   // Create node
+  auto batch_matmul = CreateBatchMatMul2(func_graph, lstm_input_grad, value_node);
   std::vector<AnfNodePtr> reduce_sum_inputs = {NewValueNode(std::make_shared<Primitive>(prim::kPrimReduceSum->name())),
-                                               lstm_input_grad};
+                                               batch_matmul};
   auto reduce_sum = func_graph->NewCNode(reduce_sum_inputs);
   // Set infer data type and shape
-  AnfAlgo::SetOutputInferTypeAndShape({AnfAlgo::GetOutputInferDataType(dynamic_rnn_grad_cnode, 1)},
-                                      {AnfAlgo::GetOutputInferShape(dynamic_rnn_grad_cnode, 1)}, reduce_sum.get());
+  auto out_shape = {AnfAlgo::GetOutputInferShape(lstm_input_grad, 0)[2]};
+  AnfAlgo::SetOutputInferTypeAndShape({kNumberTypeFloat16}, {out_shape}, reduce_sum.get());
   // Set attr
-  AnfAlgo::SetNodeAttr(kAttrAxis, MakeValue(std::vector<int64_t>{0, 1}), reduce_sum);
+  AnfAlgo::SetNodeAttr(kAttrAxis, MakeValue(std::vector<int64_t>{0}), reduce_sum);
   AnfAlgo::SetNodeAttr(kAttrKeepDims, MakeValue(false), reduce_sum);
   AnfAlgo::SetNodeAttr("is_backend_insert", MakeValue(true), reduce_sum);
   return reduce_sum;
@@ -486,8 +526,9 @@ const AnfNodePtr DynamicRnnGradFissionV2::Process(const FuncGraphPtr &func_graph
     make_tuple_inputs.emplace_back(batch_matmul);
   }
 
+  auto value_node = CreateValueNode(func_graph, dynamic_rnn_grad_cnode);
   // create reduce_sum_2
-  auto db_reduce_sum = CreateDbReduceSum(func_graph, dynamic_rnn_grad_cnode, lstm_input_grad);
+  auto db_reduce_sum = CreateDbReduceSum(func_graph, dynamic_rnn_grad_cnode, lstm_input_grad, value_node);
   make_tuple_inputs.emplace_back(db_reduce_sum);
   make_tuple_inputs.insert(make_tuple_inputs.end(), new_outputs.begin(), new_outputs.end());
   auto make_tuple = func_graph->NewCNode(make_tuple_inputs);

mindspore/ccsrc/transform/graph_ir/op_declare/rnn_declare.cc

@@ -124,7 +124,6 @@ ATTR_MAP(DynamicGRUV2Grad) = {{"direction", ATTR_DESC(direction, AnyTraits<std::
                               {"cell_clip", ATTR_DESC(cell_clip, AnyTraits<float>())},
                               {"num_proj", ATTR_DESC(num_proj, AnyTraits<int64_t>())},
                               {"time_major", ATTR_DESC(time_major, AnyTraits<bool>())},
-                              {"bias_type", ATTR_DESC(bias_type, AnyTraits<std::string>())},
                               {"gate_order", ATTR_DESC(gate_order, AnyTraits<std::string>())},
                               {"reset_after", ATTR_DESC(reset_after, AnyTraits<bool>())}};
 OUTPUT_MAP(DynamicGRUV2Grad) = {{0, OUTPUT_DESC(dw_input)},  {1, OUTPUT_DESC(dw_hidden)}, {2, OUTPUT_DESC(db_input)},

mindspore/core/ir/tensor.cc

@@ -429,6 +429,8 @@ TensorDataPtr MakeTensorData(TypeId data_type, const ShapeVector &shape, const A
       return std::make_shared<TensorDataImpl<uint64_t>>(shape, args...);
     case kNumberTypeFloat16:
       return std::make_shared<TensorDataImpl<float16>>(shape, args...);
+    case kNumberTypeFloat:
+      return std::make_shared<TensorDataImpl<float>>(shape, args...);
     case kNumberTypeFloat32:
       return std::make_shared<TensorDataImpl<float>>(shape, args...);
     case kNumberTypeFloat64:

mindspore/ops/_grad/grad_nn_ops.py

@@ -956,7 +956,7 @@ def get_bprop_dynamic_rnn(self):
 def get_bprop_dynamic_gru_v2(self):
     """Grad definition for `DynamicGRUV2` operation."""
     dynamic_gru_v2_grad = G.DynamicGRUV2Grad(self.direction, self.cell_depth, self.keep_prob, self.cell_clip,
-                                             self.num_proj, self.time_major, 'double_bias', self.gate_order,
+                                             self.num_proj, self.time_major, self.gate_order,
                                              self.reset_after)
 
     def bprop(x, winput, whidden, binput, bhidden, seq, init_h, out, dout):

mindspore/ops/_op_impl/tbe/basic_lstm_cell_c_state_grad_v2.py

@@ -38,7 +38,7 @@ basic_lstm_cell_c_state_grad_op_info_v2 = TBERegOp("BasicLSTMCellCStateGradV2")
     .output(1, "dct_1", False, "required", "all") \
     .dtype_format(DataType.F32_FracNZ, DataType.F32_FracNZ, DataType.F32_FracNZ, DataType.F32_FracNZ,
                   DataType.F32_FracNZ, DataType.F32_FracNZ, DataType.F32_FracNZ, DataType.F32_FracNZ,
-                  DataType.F32_FracNZ, DataType.F32_FracNZ, DataType.F32_FracNZ) \
+                  DataType.F32_FracNZ, DataType.F16_FracNZ, DataType.F32_FracNZ) \
     .dtype_format(DataType.F16_FracNZ, DataType.F16_FracNZ, DataType.F16_FracNZ, DataType.F16_FracNZ,
                   DataType.F16_FracNZ, DataType.F16_FracNZ, DataType.F16_FracNZ, DataType.F16_FracNZ,
                   DataType.F16_FracNZ, DataType.F16_FracNZ, DataType.F16_FracNZ) \

mindspore/ops/operations/_grad_ops.py

@@ -1344,7 +1344,6 @@ class DynamicGRUV2Grad(PrimitiveWithInfer):
         cell_clip (float): A float identifying the cell clip in the op. Default: -1.0.
         num_proj (int): An integer identifying the num proj in the op. Default: 0.
         time_major (bool): A bool identifying the time major in the op. Default: True.
-        bias_type (str): An string identifying the type of bias_type function in the op. Default to "double_bias".
         gate_order (str): An string identifying the gate order in weight and bias. Default: 'rzh.
             'zrh' is another option.
         reset_after (bool): An bool identifying whether to apply reset gate after matrix multiplication. Default: True.
@@ -1402,7 +1401,6 @@ class DynamicGRUV2Grad(PrimitiveWithInfer):
                  cell_clip=-1.0,
                  num_proj=0,
                  time_major=True,
-                 bias_type="double_bias",
                  gate_order="rzh",
                  reset_after=True):
         self.cell_depth = validator.check_value_type("cell_depth", cell_depth, [int], self.name)
@@ -1411,8 +1409,6 @@ class DynamicGRUV2Grad(PrimitiveWithInfer):
         self.num_proj = validator.check_non_negative_int(num_proj, "num_proj", self.name)
         self.time_major = validator.check_value_type("time_major", time_major, [bool], self.name)
         self.direction = validator.check_string(direction, ['UNIDIRECTIONAL'], "direction", self.name)
-        self.bias_type = validator.check_string(bias_type,
-                                                ['no_bias', 'single_bias', 'double_bias'], "bias_type", self.name)
         self.gate_order = validator.check_string(gate_order, ['zrh', 'rzh'], "gate_order", self.name)
         self.reset_after = validator.check_value_type("reset_after", reset_after, [bool], self.name)
         self.add_prim_attr("io_format", "ND")