It is, in fact, logically impossible for any observer to exactly simulate any system in which the observer is included because it would lead to infinite recursion. This is fairly straight-forward, but there is a mathematical proof out there... I honestly wish I remembered where it was because I've been looking for it for a long time.
Even if it were not logically impossible, it is practically impossible for the reasons cited by John and Alfred. Additionally, a simulation is simply a calculation that yeilds an expected value at a later time. However, The universe (as far as we can tell) follows the principle of least action meaning that the most energetically efficient method of calculating the outcome of any particular event (as would be necessary for your simulation) is the actual event itself. Meaning, the most efficient method of calculating the future state of an object subject to specified forces, is to run the experiment and measure the result.
In other words, it takes more energy for Ventner's simulation of a bacterium than the bacterium uses to go through the same motion. Practically speaking, then, the most efficient "simulation" of the universe, is the universe itself.
could we simulate the Universe With what could we possibly simulate all there is?
– Alfred Centauri Jul 31 '12 at 20:12