Enable input pointer caching in XNNPACK integration. (#42840)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42840
By caching input/output pointers and input parameters we enable the use
of caching allocator and check if we get the same input/output pointers.
If so we skip setup steps.
Test Plan:
python test/test_xnnpack_integration.py
Imported from OSS
Reviewed By: AshkanAliabadi
Differential Revision: D23044585
fbshipit-source-id: ac676cff77f264d8ccfd792d1a540c76816d5359