[pytorch] disable per-op profiling for internal mobile build (#41825)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41825
Add flag to gate D21374246 (https://github.com/pytorch/pytorch/commit/e7a09b4d17010e60bc95c25f8165ef479d1b9612) to mitigate mobile size regression.
ghstack-source-id: 108212047
Test Plan: CI
Reviewed By: linbinyu
Differential Revision: D22650708
fbshipit-source-id: ac9318af824ac31f519b7d5b4fe72df892d8d3f9
