I recently had to use a timer in a Python-program, and happened to run into a problem with the time.clock() function, which is supposed to be the most accurate [more accurate than time.time()]. The problem was, I found not the accuracy, but rather that clock() seemed to function poorly, when some blocing code, ie. socket.recv() and socket.send() was involved.

The problem:

import time
 tstart = time.clock()
 elapsed = time.clock() - tstart

The solution:

from datetime import datetimetstart = datetime.now()
 elapsed = datetime.now() - start
 elapsed.seconds

Worked just perfectly.