python-times

Get the real, user and system time used by your Python programs from your Python code.

This package implements a new type called Times to pick up the real, user and system time used for a while by one Python program. This type is a Python implementation of the bash command called time to be used by the programmer to check those times as many times as he want in running time.

Once the Times is instantiated it picks up the current values of the clock system using the times syscall to get after a while the real, user and system time spent at some specific point using the Times.times method.

As exampe:

>>> import time
>>> from times import Times
>>> t = Times()
>>> time.sleep(1)
>>> print "Real {0}s, User {1}s, Sys {2}s".format(*t.times())
Real 1.21s, User 0.0s, Sys 0.0s

Sometimes know the time consumed by our programs splitted between the real, user and system time is useful to find out where our program is really consuming the time. The following snippet of Python code finds out the cost of create 10K threads.

import threading
import times

class MyThread(threading.Thread):
    def run(self):
        # do nothing and return
        pass

t = times.Times()
threads = [MyThread() for i in xrange(10000)]
map(lambda t: t.start(), threads)
map(lambda t: t.join(), threads)
print "Real {0}s, User {1}s, Sys {2}s".format(*t.times())

The output shows that the major time is still consumed by the Python code and not by the Operating System:

$ python threads.py 
Real 0.47s, User 0.43s, Sys 0.18s

The System time and User time are the CPU-times consumed by the programs. In scenarios with many processors and many threads as the previous example, the Real time would be lesser than the sum of the System and User times

Install

Python-times comes with the minium files to be installed as a Python package build using the setuptools. Install it can be done following the following command.

$ python setup.py install

Contributing and testing

To run the tests first you should create a virtualenv where Python-times will be deployed in development way. Use the following instructions:

$ mkvirtualenv python-times
$ python setup.py develop

And then just execute the test with the following command

$ python test_times.py
.
----------------------------------------------------------------------
Ran 1 test in 1.002s

OK