Carl Friedrich Bolz-Tereick pushed to branch branch/default at PyPy / pypy
Commits: 2a61de3a by Carl Friedrich Bolz-Tereick at 2022-08-19T10:53:59+02:00 #3781: be more systematic about conditional unrolling in list, dict, tuple code. in particularly, don't do isvirtual(w_list) any more, because the list can be virtual where the underlying storage is not. --HG-- branch : list-unrolling-bugs-3781 - - - - - 1e6757a2 by Carl Friedrich Bolz-Tereick at 2022-08-19T21:53:29+02:00 #3781: jit.isvirtual would wrongly return True for freshly allocated arrays that have an unknown length. fix this --HG-- branch : list-unrolling-bugs-3781 - - - - - 6e653025 by Carl Friedrich Bolz-Tereick at 2022-08-21T13:16:04+02:00 3781: turns out isvirtual is also wrongly True for resizable lists (and that also means dicts, I suppose). Therefore require the length to be known in loop_unrolling_heuristic to work around that problem. --HG-- branch : list-unrolling-bugs-3781 - - - - - ac5c6bce by Carl Friedrich Bolz-Tereick at 2022-08-23T18:01:39+02:00 fix test_cffi_init_struct_with_list with slighly annoying code, see comment --HG-- branch : list-unrolling-bugs-3781 - - - - - 17f035ae by Carl Friedrich Bolz-Tereick at 2022-08-24T12:42:52+02:00 merge list-unrolling-bugs-3781 to fix #3781: review all the use cases of jit.loop_unrolling_heuristic, to unroll less aggressively - - - - - 9 changed files: - pypy/module/_cffi_backend/ctypestruct.py - pypy/objspace/std/dictmultiobject.py - pypy/objspace/std/listobject.py - pypy/objspace/std/specialisedtupleobject.py - pypy/objspace/std/tupleobject.py - rpython/jit/metainterp/heapcache.py - rpython/jit/metainterp/test/test_ajit.py - rpython/jit/metainterp/test/test_heapcache.py - rpython/rlib/jit.py View it on Heptapod: https://foss.heptapod.net/pypy/pypy/-/compare/2e56325b30488c74c9161502d6b708df04dc6547...17f035ae202c70a81860c7589402f78d9039612a -- View it on Heptapod: https://foss.heptapod.net/pypy/pypy/-/compare/2e56325b30488c74c9161502d6b708df04dc6547...17f035ae202c70a81860c7589402f78d9039612a You're receiving this email because of your account on foss.heptapod.net.
_______________________________________________ pypy-commit mailing list -- pypy-commit@python.org To unsubscribe send an email to pypy-commit-le...@python.org https://mail.python.org/mailman3/lists/pypy-commit.python.org/ Member address: arch...@mail-archive.com
