test=huawei_ascend_npu

PaddlePaddle · Oct 19, 2021 · 26c7b73 · 26c7b73
1 parent ebaac47
commit 26c7b73
Show file tree

Hide file tree

Showing 5 changed files with 11 additions and 30 deletions.
diff --git a/lite/backends/nnadapter/nnadapter/core/operation/range.cc b/lite/backends/nnadapter/nnadapter/core/operation/range.cc
@@ -36,9 +36,8 @@ int PrepareRange(hal::Operation* operation) {
     auto start_data = reinterpret_cast<float*>(start_operand->buffer)[0];
     auto limit_data = reinterpret_cast<float*>(limit_operand->buffer)[0];
     auto delta_data = reinterpret_cast<float*>(delta_operand->buffer)[0];
-    int64_t size = 0;
-    GetSize(start_data, limit_data, delta_data, &size);
-    output_type.dimensions.data[0] = size;
+    output_type.dimensions.data[0] =
+        GetSpanCount(start_data, limit_data, delta_data);
   } else {
     output_type.dimensions.data[0] = NNADAPTER_UNKNOWN;
   }

diff --git a/lite/backends/nnadapter/nnadapter/utility/utility.h b/lite/backends/nnadapter/nnadapter/utility/utility.h
@@ -220,10 +220,10 @@ inline int64_t GetCurrentUS() {
 }
 
 template <typename T>
-void GetSize(T start, T end, T step, int64_t* size) {
-  *size = std::is_integral<T>::value
-              ? ((std::abs(end - start) + std::abs(step) - 1) / std::abs(step))
-              : std::ceil(std::abs((end - start) / step));
+int64_t GetSpanCount(T start, T end, T step) {
+  return std::is_integral<T>::value
+             ? ((std::abs(end - start) + std::abs(step) - 1) / std::abs(step))
+             : std::ceil(std::abs((end - start) / step));
 }
 
 }  // namespace nnadapter
diff --git a/lite/core/optimizer/mir/elimination/range_calc_offline_pass.cc b/lite/core/optimizer/mir/elimination/range_calc_offline_pass.cc
@@ -29,10 +29,10 @@ namespace lite {
 namespace mir {
 
 template <typename T>
-void GetSize(T start, T end, T step, int64_t* size) {
-  *size = std::is_integral<T>::value
-              ? ((std::abs(end - start) + std::abs(step) - 1) / std::abs(step))
-              : std::ceil(std::abs((end - start) / step));
+int64_t GetSpanCount(T start, T end, T step) {
+  return std::is_integral<T>::value
+             ? ((std::abs(end - start) + std::abs(step) - 1) / std::abs(step))
+             : std::ceil(std::abs((end - start) / step));
 }
 
 void RangeCalcOfflinePass::Apply(const std::unique_ptr<SSAGraph>& graph) {
@@ -70,8 +70,7 @@ void RangeCalcOfflinePass::RemoveRangePattern(
     auto out_t = out_var->GetMutable<lite::Tensor>();
 
     // Calc range
-    int64_t size = 0;
-    GetSize(start, end, step, &size);
+    int64_t size = GetSpanCount(start, end, step);
 
     out_t->Resize(DDim({size}));
     auto out_data = out_t->mutable_data<float>();

diff --git a/lite/core/optimizer/mir/pattern_matcher.cc b/lite/core/optimizer/mir/pattern_matcher.cc
@@ -473,22 +473,6 @@ PMNode *PMNode::assert_is_op_nth_input(const std::string &op_type,
   return this;
 }
 
-PMNode *PMNode::assert_is_not_op_input(const std::string &op_type) {
-  assert_is_var();
-  asserts_.emplace_back([=](const Node *x) {
-    for (auto *op : x->outlinks) {
-      if (op && op->IsStmt()) {
-        auto *op_info = op->stmt()->op_info();
-        if (op_info->Type() == op_type) {
-          return false;
-        }
-      }
-    }
-    return true;
-  });
-  return this;
-}
-
 PMNode *PMNode::assert_is_op_output(const std::string &op_type,
                                     const std::string &argument) {
   assert_is_var();

diff --git a/lite/core/optimizer/mir/pattern_matcher.h b/lite/core/optimizer/mir/pattern_matcher.h
@@ -141,7 +141,6 @@ struct PMNode {
   PMNode* assert_is_op_nth_output(const std::string& op_type,
                                   const std::string& argument,
                                   int nth);
-  PMNode* assert_is_not_op_input(const std::string& op_type);
 
   template <typename T>
   PMNode* assert_op_attr_satisfied(