`jemalloc` future/replacement?

[owenprough-sift]

As of 2025-06-02, the jemalloc github repo is archived and no longer receiving updates/fixes. It seems like maybe there's a fork underway by Meta/Facebook (link), but this "jemalloc Postmortem" post by one of the original jemalloc developers/maintainers does not contain any details about the future of the allocator.

Because polars uses jemalloc as its memory allocator, what (if any) steps does polars need to take in order to continue being functional, secure, and generally terrific?

[coastalwhite]

Sad to see.

Honestly, we have been feeling the degradation of jemalloc already and have been thinking about what we can do next. We are not sure yet.

[skellys]

Linking this issue with another where there’s a discussion on excess memory consumption for certain use cases, with rough workarounds for jemalloc’s shortcomings: #23128

[aut0clave]

I encounter mimalloc today and wonder if it could be a suitable replacement: https://github.com/microsoft/mimalloc

[d-reynol]

Hey @ritchie46 / @orlp,

Is there any news on this front?

I'm running into issues similar to the referenced item and want to see if there's anything in the works.

[cmdlineluser]

It seems the repo has been unarchived¹.

Footnotes

1. https://engineering.fb.com/2026/03/02/data-infrastructure/investing-in-infrastructure-metas-renewed-commitment-to-jemalloc/ ↩