I'm sure there are controller/motor combinations that are less efficient than others. However, once at cruise speed, say 20 mph on up, I think you are looking to pick some high hanging fruit trying to get more range by your choices.
1%, or even a 5% difference at cruise speed is tiny, compared to the changes you can get by riding just 1 mph slower, like 19 mph vs 20. Once you get below 20 mph, then pedaling can cause as much as a 50% increase in range vs 10 mph faster.
As for full throttle efficiency, think about it, What difference would 6 fets not pwm ing make vs 12 fets not in pwm. both are going to be feeding straight current to the motor. Your controller runs at peak efficiency full throttle, but the wind will steal back 2000% of the tiny gain in controller efficiency.
On the other hand, matching your voltage, controller specs, motor, and wheel size for the particular needs of your vehicle, can easily result in 20% difference in overall efficiency, because of how it works during the start up part of the ride. I you want to dial in that acceleration part of the rides efficiency, then you need to tell us everything. The total weight of you and everything, grades, speeds you want to go, wheel size, how often you do have to stop or at least slow down, etc etc. If you have a big wheel, high speed winding, and not enough wattage, then bumping watts with a voltage increase can help. Quicker you get to speed, the sooner your amps draw drops. five seconds at 1000w is less watt hours than 30 seconds at 500w. But it still will not be more efficient than pedaling to 10 mph, then grabbing throttle.