If core icing did occur, then the fuel most probably was no longer homogenous, particularly if wax was floating on the surface.
Most probably everything worked as advertised, but the heavier parts of the fuel(and water) did not possess enough heat(kJ) to provide the required thrust, despite the amount of fuel in all probability being correct for fuel in its normal condition.
Not sure, but I estimate that the fuel delivery pipe would be around 4" dia to give enough fuel for max. power, so ice or anything else would not restrict flow, particularly as the fuel is under pressure.
My final 2 cents worth is that evidence of cavitation on the pumps is only significant in that it takes quite some time for cavitation to become evident on the impellor surface, so most unlikely to have been a contributor to this problem.
Cheers