The timeit command#

This uses the function sage_timeit().

class sage.misc.sage_timeit_class.SageTimeit#

Bases: object

Time execution of a command or block of commands.

Displays the best WALL TIME for execution of the given code. This is based on the Python timeit module, which avoids a number of common traps for measuring execution times. It is also based on IPython’s %timeit command.


timeit(statement, preparse=None, number=0, repeat=3, precision=3)


sage: timeit('2^10000')
625 loops, best of 3: ... per loop

We illustrate some options:

sage: timeit('2+2',precision=2,number=20,repeat=5)
20 loops, best of 5: ... per loop

The preparser is on by default (if it is on), but the preparse option allows us to override it:

sage: timeit('2^10000', preparse=False, number=50)
50 loops, best of 3: ... per loop

The input can contain newlines:

sage: timeit("a = 2\nb=131\nfactor(a^b-1)", number=25)                          # needs sage.libs.pari
25 loops, best of 3: ... per loop

See also


eval(code, globs=None, locals=None, **kwds)#

This eval function is called when doing %timeit in the notebook.


  • code – string of code to evaluate; may contain newlines.

  • globs – global variables; if not given, uses module scope globals.

  • locals – ignored completely.

  • kwds – passed onto sage_timeit. Common options are preparse, number, repeat, precision. See sage_timeit() for details.

OUTPUT: string – timing information as a string


sage: timeit.eval("2+2")                     # random output
'625 loops, best of 3: 1.47 us per loop'

We emphasize that timeit times WALL TIME. This is good in the context of Sage where commands often call out to other subprocesses that do not appear in CPU time.

sage: timeit('sleep(float(0.5))', number=3)  # long time (5s on sage.math, 2012)
3 loops, best of 3: ... ms per loop