Unless you are hitting large binary checkins and not utilizing LFS I can't imagine a scenario where another current RCS would perform better.

"Large binary checkins" is not the actual issue. Git degrades as repository size increases. Large binary checkins make it much easier / faster to reach this situation, but you can also reach it just fine with regular text-based repository if they're big and have a significant history (long-lived and many contributors).

Unless you mean another DVCS, P4 can run circles around git on large repos (mostly on account of not conceptually trying to copy the entire state of the entire repo at every commit).

Run a reconcile offline changes on an actually large repo and come back and tell me that again. ;)