PyCon AU 2010: Faster Python Code and Memory Consumption resources

Making Your Python Code Fast

Tools mentioned in my talk:

timeit — Microbenchmarking module. http://docs.python.org/library/timeit.html
kcachegrind — GUI for “calltree” output from lsprof/cProfile. http://kcachegrind.sourceforge.net/html/Home.html
strace — System call tracer for Linux. http://sourceforge.net/projects/strace/
perf — Performance analysis tools for Linux. apt-get install linux-tools
bzr — Take a look at the bzrlib.profile_imports, bzrlib.breakin and bzrlib.lsprof modules. http://bazaar.canonical.com/
netem — Add 500ms latency to loopback network. sudo tc qdisc add dev lo root netem delay 500ms
tcptrace — Analyse TCP traffic dumps. http://www.tcptrace.org/
cpufreq-set — Force CPU to full speed for benchmarking. cpufreq-set -g performance

You can watch or download the video.

Tools mentioned in my talk:

Handy Python builtins — sys.getsizeof, sys.getrefcount, gc.get_referrers, and gc.get_objects.
Meliae — Analyse memory consumption of a Python program. https://launchpad.net/meliae
Heapy — Another memory analyser. http://guppy-pe.sourceforge.net/#Heapy
objgraph — Draw diagrams of object reference graphs. http://mg.pov.lt/objgraph/
valgrind — Find sources of leaks and crashes in C code. (Follow the instructions in Python's README.valgrind). http://www.valgrind.org/docs/manual/index.html

You can watch or download the video.