Skip to content

Add note [Note: Re-using upstream TensorImpl] #4284

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Dec 9, 2022
Merged

Add note [Note: Re-using upstream TensorImpl] #4284

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
023e4d5
Add note [Note: Re-using upstream TensorImpl]
wonjoo-wj Dec 9, 2022
14e9363
Run linter
wonjoo-wj Dec 9, 2022
20b4481
Remove test file
wonjoo-wj Dec 9, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions torch_xla/csrc/tensor_impl.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,15 @@

namespace torch_xla {

// [Note: Re-using upstream TensorImpl] As part of the LTC migration effort,
// we tried to re-use the upstream LTCTensorImpl
// (torch/csrc/lazy/core/tensor_impl.h) instead of having our own version of
// XLATensorImpl. However, LTCTensorImpl defines its own set of hard-coded
// dispatch keys in its constructor and has functions that call this
// constructor. In addition, updating XLATensorImpl to extend LTCTensorImpl does
// not produce much benefit since that wouldn't remove much duplicate code. As a
// result, we decided to keep and use XLATensorImpl.

// Tensor implementation class used to be fed to the at::Tensor.
// Its scope is just to handle an XLATensor.
class XLATensorImpl : public c10::TensorImpl {
Expand Down