Don't make factory methods create a tensor and then immediately copy it (#17565)

Summary:
Create a `make_variable` override that moves out of a tensor instead of going through `shallow_copy_and_detach`. Call this override from factory methods like `empty` that create a brand new tensor, do nothing with it, and then copy it into a variable.

Will update this with actual numbers, but it seems to get rid of around 20-40% of the overhead of calling `torch.empty(0)`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17565

Differential Revision: D14266130

Pulled By: umanwizard

fbshipit-source-id: f57d5f2ca3f80ee8ee96d50f905e852fd10db941

diff --git a/tools/autograd/gen_variable_factories.py b/tools/autograd/gen_variable_factories.py
index b193fcc..fea70a3 100644 (file)
--- a/tools/autograd/gen_variable_factories.py
+++ b/tools/autograd/gen_variable_factories.py
@@ -13,7 +13,7 @@ inline at::Tensor ${name}(${formals}) {
   ${pre_record_trace}
   at::Tensor tensor = at::${name}(${actuals});
   at::Tensor result =
-    autograd::make_variable(tensor, /*requires_grad=*/${requires_grad});
+    autograd::make_variable_consuming(std::move(tensor), /*requires_grad=*/${requires_grad});
   ${post_record_trace}
   return result;
 }

diff --git a/torch/csrc/autograd/variable.h b/torch/csrc/autograd/variable.h
index 7eea2b0..e726f9c 100644 (file)
--- a/torch/csrc/autograd/variable.h
+++ b/torch/csrc/autograd/variable.h
@@ -104,7 +104,8 @@ struct TORCH_API Variable : public at::Tensor {
       bool allow_tensor_metadata_change,
       Edge gradient_edge);
 
-  /// Creates a `Variable` from the given `Tensor`. `requires_grad` should be
+  /// Creates a `Variable` from the given `Tensor`, copying its underlying `TensorImpl`.
+  /// `requires_grad` should be
   /// set only for leaves, and determines whether the `Variable` will accumulate
   /// gradients. NOTE: `data` must *not* be a `Variable` already. Its dynamic
   /// type *must* be `Tensor`.
@@ -113,6 +114,16 @@ struct TORCH_API Variable : public at::Tensor {
       bool requires_grad,
       bool allow_tensor_metadata_change);
 
+  /// Creates a `Variable` from the given `Tensor`, consuming its underlying `TensorImpl`.
+  /// This is intended to be used from functions that immediately create a `Tensor`,
+  /// convert it to a `Variable`, and then free it; it has been found to
+  /// decrease the overhead of those operations, in some situations.
+  /// The comments about `requires_grad` and `data` on the above version also apply to this one.
+  friend Variable make_variable_consuming(
+      at::Tensor data,
+      bool requires_grad,
+      bool allow_tensor_metadata_change);
+
   /// Creates a `Variable` from the given `Tensor` and specify a
   /// `gradient_edge`, i.e. a (function, input_nr) pair specifying the function
   /// in the autograd graph, and what particular input of that function, this
@@ -576,6 +587,22 @@ inline Variable make_variable(
   return Variable();
 }
 
+inline Variable make_variable_consuming(
+    at::Tensor data,
+    bool requires_grad = false,
+    bool allow_tensor_metadata_change = true) {
+  AT_CHECK(
+      !data.is_variable(),
+      "Must not create a new variable from a variable, use its .data()");
+  if (data.defined()) {
+    AT_ASSERT(data.getIntrusivePtr().use_count() == 1);
+    data.unsafeGetTensorImpl()->set_allow_tensor_metadata_change(allow_tensor_metadata_change);
+    auto autograd_meta = c10::guts::make_unique<Variable::AutogradMeta>();
+    return Variable(c10::make_intrusive<Variable::Impl>(std::move(data), std::move(autograd_meta), requires_grad));
+  }
+  return Variable();
+}
+
 inline Variable make_variable(
     at::Tensor data,
     Edge gradient_edge,