On Tue, Jan 13, 2009 at 5:09 PM, Marc Tompkins <marc.tompk...@gmail.com> wrote:
> Apparently nothing at all is wrong with it: > C:\Python25\Lib>python timeit.py -s "import math" "x=math.exp(10)" > 1000000 loops, best of 3: 0.678 usec per loop > > C:\Python25\Lib>python timeit.py -s "from math import e" "x=e**10" > 1000000 loops, best of 3: 0.411 usec per loop Using ** beats exp() because there is an explicit bytecode for exponentiation so no function call is needed. Careful, though; on my computer about 1/2 of the difference between these two is due to the different style of import: C:\Project\Mango> python -m timeit -s "import math" "x=math.exp(10)" 1000000 loops, best of 3: 0.467 usec per loop C:\Project\Mango> python -m timeit -s "from math import exp" "x=exp(10)" 1000000 loops, best of 3: 0.352 usec per loop C:\Project\Mango> python -m timeit -s "from math import e" "x=e**10" 1000000 loops, best of 3: 0.259 usec per loop Name lookup and function call are both expensive in Python. Using a constant instead of the variable e is substantially faster: C:\Project\Mango> python -m timeit "x=2.718182**10" 10000000 loops, best of 3: 0.0399 usec per loop Kent _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor
