New submission from Dennis Sweeney <sweeney.dennis...@gmail.com>:
See the issue here for lots of data from before specializing: https://github.com/faster-cpython/ideas/issues/105 See https://gist.github.com/sweeneyde/91855e50feb9992b604ddda2d4f1511e for specialization data, pyperformance benchmarks, and microbenchmarks. Right now the PR draft specializes list[int], dict[str], and bytearray[int], although I don't if the bytearray opcode should be kept. I considered specializing dict[object], but I figured the overhead of PyDict_SetItem isn't that much less than PyObject_SetItem. ---------- components: Interpreter Core messages: 405015 nosy: Dennis Sweeney priority: normal severity: normal status: open title: Specialize STORE_SUBSCR type: performance versions: Python 3.11 _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue45609> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
