René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Closed figures linger in memory","articleBody":"tl;dr: When I repeatedly create a large figure, save it, and close it, memory usage keeps growing.\r\n\r\nOver at [this discussion about when MPL should trigger garbage collection](https://github.com/matplotlib/matplotlib/pull/3045), @efiring had some lingering doubts about the chosen solution:\r\n\r\n\u003e It would certainly be good to have a clearer understanding of when, if ever in practice, it would lead to troublesome increases in memory consumption\r\n\r\nI ran into such a case today, when my batch job filled up 60G of RAM over night. \r\n\r\nI repeatedly create a large figure, save it, then close it. If I don't manually call `gc.collect()` after closing each figure, memory consumption saturates at around 10x of what an individual figure needs. In my case, with several fairly complex figures, this was enough to fill a big machine.\r\n\r\nSince this is not obvious from the docs, I think there should be an official way to go back to more aggressive GC for cases like this where the trade-off discussed at https://github.com/matplotlib/matplotlib/pull/3045 fails. Maybe `close(force_gc=True)`?\r\n\r\n**Code for reproduction**\r\n\r\n![image](https://cloud.githubusercontent.com/assets/486707/25231298/a301d630-25d7-11e7-9592-406d67192088.png)\r\n\r\n```python\r\nfrom memory_profiler import profile # https://pypi.python.org/pypi/memory_profiler\r\nfrom memory_profiler import memory_usage\r\n\r\nimport matplotlib.pyplot as plt\r\nimport numpy as np\r\nimport gc\r\n\r\nN = 80\r\n\r\n@profile\r\ndef do_plots():\r\n    fig = plt.figure()\r\n    plt.plot(np.random.rand(50000))\r\n    plt.savefig('/tmp/bla.png')\r\n    plt.close(fig)\r\n\r\ndef default():\r\n    for k in range(N):\r\n        print(k)\r\n        do_plots()\r\n\r\ndef manual_gc():\r\n    for k in range(N):\r\n        print(k)\r\n        do_plots()\r\n        gc.collect()\r\n\r\n\r\nmem_manual_gc = memory_usage((manual_gc, [], {}))\r\nmem_default = memory_usage((default, [], {}))\r\n\r\n\r\nplt.plot(mem_manual_gc, label='gc.collect() after close')\r\nplt.plot(mem_default, label='default behaviour')\r\nplt.ylabel('MB')\r\nplt.xlabel('time (in s * 0.1)')  # `memory_usage` logs every 100ms\r\nplt.legend()\r\nplt.title('memory usage')\r\nplt.show()\r\n```\r\n\r\n**Matplotlib version**\r\n  * Operating System: Ubuntu 16.10\r\n  * Matplotlib Version: 2.0.0\r\n  * Python Version: 3.5.2\r\n","author":{"url":"https://github.com/cknd","@type":"Person","name":"cknd"},"datePublished":"2017-04-20T13:10:59.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":28},"url":"https://github.com/8519/matplotlib/issues/8519"}