I ran some numbers in the following scenario: examine the range of needing to roll a 3, a 4, a 5... up to a 19 to hit the enemy while wielding a +3 weapon (thus a 4...20 for a +2 weapon). I originally ran this with just a "core" range of 5-15 (6-16) and found that the GA > Maul > GS, but a 1d12 GS > ALL, until you factor in expanded crit and then GA is back on top. So I decided to compare the entire range minus the absurd edge cases like where the GS actually crits on 20 but the GA is only just auto-hitting. To account for the edge cases being, well, edge cases, I multiplied the 4 lowest and 4 highest rolls needed damage values by .2, .4, .6, and .8 in order of distance from the center.
Then the average damage per swing is:
Level 1 test, 1[W]+ability (at +4):
(21-number needed to roll)/20*(avg weapon dmg + ability) + 0.05*(
(max weapon dmg - avg weapon dmg) + avg high crit dmg per [W])
Level 28 test, 2[W]+ability (at +8):
(21-roll needed)/20*(2*avg wpn dmg+ability)+0.1*(2*(max wpn - avg wpn)+3*high crit avg per [W])
Note that at the end for the Critical hit which happens with p = 0.05 or 1/20 times, we calculate damage by subtracting the average weapon damage from the max damage because we already counted the weapon average damage once on this hit in the previous part. We then add the weapon's high crit damage, if any. This effectively gives you the extra damage on a crit. For example, a longsword has average damage 4.5 and max damage 8 and no high crit. So if you need to roll an 11 to hit, then (21-11) = 10, divided by 20 = 0.5, which is the chance to hit. So we multiply that by the average damage and get 0.5*4.5. Now 1/20 of these rolls are crits, so we need to add not 8, because we already counted 4.5, but (8-4.5).
I then added up the numbers for each of these ACs for greatsword RAW, greataxe RAW, maul RAW, and greatsword 1d12, scaling as mentioned for edge cases.
No surprise, the GreatAxe beats the Maul which beats the RAW Greatsword. The RAW Greatsword is actaully only better in DPS then the GA when the greatsword needs to roll a 19, and then only at Level 1. The level 28 test is no contest at all although the RAW GS is better than the Maul when the GS wielder hits on 15+ (this is the same as at Level 1).
A 1d12 greatsword, no surprise, is a great Level 1 weapon. It only slightly edges out the GreatAxe even at this level, but is consistently better on every swing, even with the high crit. At level 28, the greataxe is significantly higher than either the maul or the 1d12 greatsword due to high crit being 3[W].
If you narrow the range and only consider where the GS needs 5-15 to hit and count them all evenly, the GS still lags behind. The 1d12 GS is only the slighest bit higher than a GA at level 1. The 1d12 GS and Maul are very close at Level 28 (the larger overall damage numbers means that the relatively fixed distance between them means they are "closer" than they were at level 1) and both trail the GA.
Interestingly, a 1d10 high crit greatsword comes out behind the greataxe even at level 1. It is very very close to the Maul at this level. At level 28, the greataxe is still ahead, but the 1d10 high crit greatsword beats the maul.
Even when I adjusted the scaling factors such that the GS needs a 2 = 0.1x, the GS needs a 3 is 0.2x ... up to the GS needs an 11 or higher is 1x, the GA *STILL* comes out ahead although it is very close.
Conclusion: 1d10 greatsword is weaker in raw damage output. High crit weapons at epic tier are REALLY good. A 1d10 high crit greatsword is behind in raw damage but is much closer. The Maul is uninteresting in most cases.