René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"pickletools optimize performance degraded with factor 3-4 in python 3.14","articleBody":"# Bug report\n\n### Bug description:\n\nThe pickletools.optimize code suddenly performs bad starting with python 3.14, with or without JIT.\nThe degradation is a factor 3 to 4.\n\nIt is easy to test.\n\nFirst use this script to make a sizeable dict and pickle it and write it to file.\n\n```python\n\nimport pickle\n\nx = {}\n\nfor i in range(1, 1000000):\n    x[i] = f\"ii{i:\u003e07}\"\n\nwith open(\"pickled\", \"wb\") as fh:\n    pickle.dump(x, fh, protocol=4)\n```\n\nThen write a script, let's call it `optpickle.py` to read the pickled dict from file and optimize it and write the result to another file.\n\n```python\n# optpickle.py\n\nimport pickletools\n\nwith open(\"pickled\", \"rb\") as fh:\n    p = fh.read()\n\ns = pickletools.optimize(p)\n\nwith open(\"pickledopt\", \"wb\") as fh:\n    fh.write(s)\n```\n\nNow execute this in three separate terminal windows, one geared to python 3.13, one to python 3.14.0 without JIT, and one to python 3.14.0 with JIT.\n\nMy results on a mac with M1 processor were\n\nversion | time elapsed | cpu time\n--- | --- | ---\n3.13 | 1.67| 1.773\n3.14.0 no JIT | 6.52 | 6.664\n3.14.0 JIT | 6.52 | 6.642\n\nThat is a dramatic drop in performance!\n\nThe filesizes are 16.9 MB for the unoptimized file and 15.9 for the optimized one.\n\nI decided to drop the optimization in an [app where I used it heavily](https://github.com/annotation/text-fabric/blob/7de3664b7ddd0ec0dc4c1c6d5da9bf04b9848b73/tf/core/data.py#L687-L688)\n\n\n### CPython versions tested on:\n\n3.14\n\n### Operating systems tested on:\n\nmacOS","author":{"url":"https://github.com/dirkroorda","@type":"Person","name":"dirkroorda"},"datePublished":"2025-10-15T18:23:12.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":16},"url":"https://github.com/140175/cpython/issues/140175"}