[dtensor][10/n] switch pointwise op to use op strategy (#111234)
As titled, this also handles sth like [Shard(0), Shard(0)] correctly for
pointwise ops, which was previously errored out
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111234
Approved by: https://github.com/fduwjj
ghstack dependencies: #109145, #110717