Armin Rigo <[email protected]> added the comment: It looks like it's not a bug or a leak, but just a very very slow JIT start-up. By running the process for long enough, RAM usage grows slowly from ~30MB to 89MB ("Resident size", Linux32), and at the same time the speed improves slowly by a total factor of roughly 3x. It takes maybe 10 minutes(!) to reach the maximum RAM usage and top speed. Afterwards, the memory usage doesn't grow past 89MB, even if I run it for one hour.
The extra ~60MB is taken by JIT code and associated data structures. I don't exactly know why it is so big in this case. But note also that its size should not depend on the quantity of real data that a real web server is handling, so 60MB is not that bad. ---------- nosy: +arigo status: unread -> chatting ________________________________________ PyPy bug tracker <[email protected]> <https://bugs.pypy.org/issue1257> ________________________________________ _______________________________________________ pypy-issue mailing list [email protected] http://mail.python.org/mailman/listinfo/pypy-issue
