performance of dictionary lookup vs. object attributes

[Fredrik Lundh]

Peter Otten wrote:

> $ python -m timeit -s'class A(object): pass' -s 'a = A(); a.alpha = 42'
> 'a.__getattribute__("alpha")'
> 1000000 loops, best of 3: 0.674 usec per loop

also:

    timeit -s "class A(object): pass" -s "a = A(); a.alpha = 42"
    "a.__getattribute__('alpha')"
    1000000 loops, best of 3: 0.592 usec per loop

    timeit -s "class A(object): pass" -s "a = A(); a.alpha = 42"
    "getattr(a, 'alpha')"
    1000000 loops, best of 3: 0.403 usec per loop

which is better, but nowhere close to

    timeit -s"class A(object): pass" -s "a = A(); a.alpha = 42" "a.alpha"
    10000000 loops, best of 3: 0.158 usec per loop

    timeit -s"class A(object): pass" -s "a = A(); a.alpha = 42; d=vars(a)"
    "d['alpha']"
    10000000 loops, best of 3: 0.126 usec per loop

on the other hand, when we're talking about sub-microsecond operations, you'll
see "massive slowdowns" as soon as you're actually trying to do something with
the data you just fetched...

</F>