Armin Rigo <[email protected]> added the comment: This is actually a problem in the richards benchmark itself, which doesn't correctly reset the state between each iteration. Fixed in c1d1f4dd5f5f. I only changed rpython/translator/goal/richards.py, which is not (I believe) the one run on the nightly benchmarks; the latter should not be modified.
---------- nosy: +arigo status: unread -> resolved ________________________________________ PyPy bug tracker <[email protected]> <https://bugs.pypy.org/issue1747> ________________________________________ _______________________________________________ pypy-issue mailing list [email protected] https://mail.python.org/mailman/listinfo/pypy-issue
