Timing cached functions

edit

You could include an optional argument which clears the function's cache:

@cached_function def fibo(n, clear_cache=False):
    if clear_cache:
        fibo.clear_cache()
    if n==0 or n==1:
        return 1
    return fibo(n-1, clear_cache)+fibo(n-2, clear_cache)

You could modify this to only remove certain inputs from the cache (just remove certain keys from fibo.cache).

Also, you can use %time instead of timeit, or tell timeit to just run once:

sage: timeit('fibo(12, True)',number=1, repeat=1)
1 loops, best of 1: 4.69 ms per loop

sage: timeit('fibo(12, False)',number=1, repeat=1)
1 loops, best of 1: 192 µs per loop

But is this really any better than @John Palmieri's suggestion?

Expanding on what niles said, you could have one function which is not cached and another that is. Run timeit on each one in the same Sage session.

If you have a method which is cached, then you can get around that by doing

sage: a = matrix(...)
# this will be fast because it calls the cached version:
sage: timeit('a.my_great_cached_method()')
# this won't be as fast, because the cache gets reset
# for each instance of the class:
sage: timeit('matrix(...).my_great_cached_method()')

There is no built-in way to tell timeit to ignore the cache, though.