On Tue, May 16, 2017 at 11:28:02AM -0700, Steve Fink wrote: > But I'm curious if you know anything about the intended future directions of > jemalloc, if there are any, and whether they align with anything we need.
As Nick mentions, Jason Evans, the main developer on jemalloc (guess where "je" comes from?) has been working at Facebook for a few years, and Facebook needs is what mainly drives jemalloc development nowadays. There are a few non-Facebook contributors, though. Theoretically, many of those changes would benefit us. But in practice, we've had many surprises. If the current jemalloc was on par with mozjemalloc wrt performance and memory usage, we may have more easily tracked and been able to influence its development, but we've been playing catch-up and failed to make it suit our needs, so that doesn't really help. > Also, how much of the difficulty in maintaining jemalloc perf across > upgrades is specific to ARM? Not much, in fact. We've had regressions on all platforms, not necessarily at the same time. Mike _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
