In a single sentence: if you don't use feats, fighters get overshadowed by the damage-dealing abilities of some cantrip-users; if you does use feats, a couple of dominant archetypes (GW, SS) crowd out the rest.
Quite a way upthread someone offered up some sorcerer numbers. I can't remember who it was, and haven't gone back for a search, but here are some quick numbers of my own:
A 6th level Dragon Sorcerer has 6 SP + 19 spell levels (4, 3, 2) for 25 SPs total. That's enough for 25 twinned cantrips, or about 6 combat's worth. Fire Bolt does 2d10+4 (assuming an 18 stat). So that's 4d10+8, or 30 damage, spread across two targets, per round, prior to factoring in the chance to hit.
The featless fighter of the same level gets 2 attacks for (say) 2d6+6 (20 stat, +1 weapon). That's 4d6+12, or 26 damage, per round. But it can be focused. Factoring in GWF style takes it from 3.5 to 25/6 per die, or 100/6 +12 = not quite 29 per round. Assuming every 2 encounters yields a long rest, and 4 rounds per encounter then 1 in 8 rounds has an Action Surge, which is another +3.5-ish damage for 32 to 33 expected damage. The fighter's chance to hit is also better than the sorcerers (+1 weapon, +1 stat for +9 rather than +7) - against AC 15, that is a 75% rather than 65% chance to hit, which is about another +6 to hit, for around 38 damage.
(If there's other stuff I'm missing, please point it out. I haven't factored in criticals, which favour the fighter, especially if a Champion.)
The fighter is about 25% ahead in damage by my maths. The fighter will also have better AC (17 or 18 at least, I'd assume, compared to 13+ DEX for 15 or so for the sorcerer). And hp at d10 rather than d6+1, plus second wind. But the fighter will be in melee, and so will have more need of these things!
The sorcerer is at range, and so doesn't have to close. This boosts damage per encounter. The sorcerer is at range, and so is more likely to suffer cover penalties. This reduces damage. I haven't tried to factor those things in.
The sorcerer has the option, at any point up to the last couple of rounds, to stop doing damage and instead use some other spell. That is a flexibility the fighter can't match. How much is it worth? I don't know, and there's probably no table-invariant measure, but surely quite a bit!
Anyway, writing that up prompted me to search back and find the post upthread that I mentioned above. Here it is:
To answer Krachek's question, I don't know if it's broken but it seems pretty strong: matching the fighter's melee damage output with quickened cantrips, and still having the option to stop doing that at any time and use its spells for other stuff instead.
Right, so this is getting into the sort of discussion I suggested upthread: can the claim that the fighter's main contribution is DPR be contested?
In my AD&D experience, the role of the fighter changed with level. At low and even mid levels, the fighter was a bold warrior who led the attack, cut down foes, and supported the MUs. But name level or thereabouts, the MUs dominated combat and the role of the fighters was to provide the sort of defence you describe, and to mop up whatever was left over after the spell assaults. I think different players have different reactions to that sort of change of role.
Anyway, if a fighter player can embrace that approach to the class, having a sorcerer compete in DPR while shielded by the fighter may not be a problem at that table.