Grassroots Motorsports recently tested 87 E10, 93 E10, straight 93, 100, 105, E85 and M1 Methanol on a Miata with a standalone computer and tuned each on a Dynapack dyno to make the most power with the fuel. The differences results were not dramatic except for M1 Methanol:
87 E10 made 135 hp and 117 lbs tq
93 E10 made 136 hp and 124 lbs tq
93 made 134 hp and 122 lbs tq
E85 made 143 hp and 128 lbs tq
100 made 137 hp and 123 lbs tq
105 made 137 hp and 124 lbs tq
M1 made 160 hp and 140 lbs tq
December 2012 GRM
At least on the Miata, E85 offers about 6% hp and 3% tq over 93 E10; race gas offers about 2% power and no torque over 93 E10, and M1 offers about 16% hp and about 11% tq over 93 E10. There seem to be just a couple of race gas tests in the forum on the E9xM3:
Pencilgeek ran gained about 7.5% on 100 versus 91, with no tune, about a 7.5% difference.
Someone else ran race gas versus pump, with no tune, and gained 2% power but lost 3% tq:
Vishnu tested meth injection on top of 91, with just ignition timing changes, and gained about 6.5% power and torque:
Race gas or E85 or M1 differences may be greater on a very high compression motor like the S65 that is probably often pulling timing and losing power relative to its rating in normal driving and should be greater on a forced induction motor with an even higher effective compression ratio under boost.