[inductor] Remove ReinterpretView copy_ for AOT Inductor outputs (#106564)
Running benchmark on HF models result in 71% pass rate now: P802905571
Updated [dashboard](https://hud.pytorch.org/benchmark/compilers?startTime=Fri%2C%2028%20Jul%202023%2005%3A02%3A20%20GMT&stopTime=Fri%2C%2004%20Aug%202023%2005%3A02%3A20%20GMT&granularity=hour&suite=torchbench&mode=inference&dtype=bfloat16&lBranch=angelayi/bench&lCommit=e35a655e59b2038c0395f972a1f567f862093d9c&rBranch=main&rCommit=3e5a52cedd2d586fc6cb40a73a098252b9edc2a1)
Originally, a lot of the HF export-aot-inductor tests are failing with the error message:
```
RuntimeError: unsupported operation: some elements of the input tensor and the written-to tensor refer to a single memory location. Please clone() the tensor before performing the operation.
```
I looked at the result of one of the models, AlbertForMaskedLM, and the error is due to an additional [`copy_`](https://www.internalfb.com/phabricator/paste/view/P802043305?lines=1460%2C1426%2C1438%2C1451%2C1428) being inserted at the end. Looking at the [exported graph](https://www.internalfb.com/phabricator/paste/view/P802908243?lines=1124), `buf237` in the cpp program corresponds to the `view_269` node. During inductor lowering, this `view_269` node will result in a `ir.ReinterpretView` node, and when generating code for the outputs, this [line](https://fburl.com/code/epola0di) will add an additional `copy_`.
I'm unsure if removing this case will result in other errors, but it seems to raise the HF model benchmark pass rate :)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/106564
Approved by: https://github.com/jansel
