There are various tools to profile (measure) performance and identify problems. * cProfile * gprof2dot * sysprof * flameprof # Profiling with cProfile Python's [cProfile](http://docs.python.org/library/profile.html) allows profiling the whole GTG app. Do this following: ./launch.sh -p 'python3 -m cProfile -o gtg.prof' Let GTG launch. Quit, and do the following to parse the results: $ ipython In [1]: import pstats In [2]: p = pstats.Stats('gtg.prof') In [3]: p.strip_dirs().sort_stats("cumulative").print_stats(20) This should display profiling results, sorted by cumulative time, and displaying the top 20 contributors. Many others sorting possibilities are available, look at the [python documentation](http://docs.python.org/library/profile.html) to learn more about it. Here's an example of output with the above sorting configuration: ``` Thu Aug 6 09:35:55 2009 gtg.prof 453156 function calls (445719 primitive calls) in 3.799 CPU seconds Ordered by: cumulative time List reduced from 1197 to 20 due to restriction <20> ncalls tottime percall cumtime percall filename:lineno(function) 1 0.000 0.000 3.802 3.802 :1() 1 0.000 0.000 3.802 3.802 {execfile} 1 0.000 0.000 3.801 3.801 gtg:28() 1 0.000 0.000 3.405 3.405 gtg.py:93(main) 1 0.000 0.000 2.599 2.599 browser.py:1405(main) 1 0.943 0.943 2.427 2.427 {gtk._gtk.main} 142 0.003 0.000 1.283 0.009 browser.py:1320(on_task_added) 961 0.009 0.000 0.917 0.001 tagtree.py:65(on_get_value) 2056 0.060 0.000 0.911 0.000 {method 'get_value' of 'gtk.TreeModel' objects} 200 0.002 0.000 0.892 0.004 requester.py:156(get_active_tasks_list) 200 0.434 0.002 0.890 0.004 requester.py:96(get_tasks_list) 142 0.003 0.000 0.888 0.006 tagtree.py:36(update_tags_for_task) 142 0.017 0.000 0.870 0.006 {method 'row_changed' of 'gtk.TreeModel' objects} 175 0.004 0.000 0.808 0.005 browser.py:796(tag_visible_func) 142 0.009 0.000 0.330 0.002 tasktree.py:197(add_task) 79 0.010 0.000 0.324 0.004 cleanxml.py:93(savexml) 1 0.002 0.002 0.286 0.286 gtg.py:46() 79 0.001 0.000 0.274 0.003 minidom.py:47(toprettyxml) 2 0.000 0.000 0.273 0.137 __init__.py:148(save_datastore) 1 0.000 0.000 0.272 0.272 __init__.py:81(get_backends_list) ``` # Graphical profiling charts with gprof2dot Install [gprof2dot](https://github.com/jrfonseca/gprof2dot), then execute: ./launch.sh -p 'python3 -m cProfile -o gtg.prof' python gprof2dot.py -f pstats gtg.prof | dot -Tpng -o output.png ...and watch the resulting pretty image! ![Generated image](https://wiki.gnome.org/Apps/GTG/development?action=AttachFile&do=get&target=profile.png) # Sysprof Sysprof is a really cool graphical user interface for system-wide (or application-specific) profiling. If it can be useful for profiling GTG, someone should document how to use it here... # flameprof (flamegraph) You can use [flameprof](https://pypi.org/project/flameprof/) to generate an [flamegraph](https://www.brendangregg.com/flamegraphs.html), which roughly shows what GTG does over time. ```sh ./launch.sh -p 'python3 -m cProfile -o gtg.prof' flameprof -o gtg.svg gtg.prof ``` ![Generated image (not GTG)](https://raw.githubusercontent.com/brendangregg/FlameGraph/master/example-perf.svg)