[PyTorch] Use a better hash table in CUDACachingAllocator (#71667)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/71667
We have flat_hash_set because it performs better than std::unordered_set.
ghstack-source-id: 148013648
Reviewed By: ngimel
Differential Revision: D33720595
fbshipit-source-id: aa6077c474dd6fc61ce17e24ebde4056c8bae361
(cherry picked from commit 386082eaf1d4669c7967ba9cdf765d9d677f5cd9)