We haven't. Google infrastructure uses tcmalloc. Is there a reason to believe it offers a significant win?

I'd expect similar performance but less fragmentation?, less memory used by the process if you aren't regularly calling MallocExtension::instance()->ReleaseFreeMemory() as a tcmalloc user.

The first answer at https://www.quora.com/Is-tcmalloc-stable-enough-for-producti... (by Keith Adams) is completely consistent with what I've seen. Rust went with jemalloc for some reason too.

IIRC jemalloc is somewhat better about releasing memory in a timely fashion, at least by default.