Temeraire: Hugepage-Aware Allocator | tcmalloc

Notes on the name1: the french word for “reckless” or “rash” :), and also the name of several large and powerful English warships. So: giant and powerful, but maybe a little dangerous. :)

This is a description of the design of the Hugepage-Aware Allocator. We have also published “Beyond malloc efficiency to fleet efficiency: a hugepage-aware memory allocator” at OSDI 2021, which provides further details on the design, implementation, and rollout of Temeraire.

reasonable allocation speed. This is really stating a non-goal: speed parity with PageHeap::New. PageHeap is a relatively light consumer of cycles. We are willing to accept a speed hit in actual page allocation in exchange for better hugepage usage and space overhead. This is not free but we think is well justified. Our goal is more to avoid catastrophic regressions in speed. We intentionally accept two particular time hits:

The algorithm – as usual here, really, the data structures, which neatly determine our algorithm – are nicely divided into components. Essentially, the path of an allocation goes like this:

Leave a Comment