From what I understand of the different stepping for controllers is say full steps moves 1/2" then half step would be 1/4" and so on. Dont know about the torque though. I think the higher the rpm the more effect it has on torque (higher rpm=less torque). I dont know about if the step has a direct effect on torque besides the rpm or not. My board only has full and half step abilitys and here is what Hoss told me when I asked a question on my controller.
"A 200 step/rev stepper (1.8 degree per step, pretty typical)
A .200 pitch ballscrew (5 revs per inch)
200 x 5 = 1000 steps/inch (full step mode)
1000 x 2 = 2000 steps/inch (1/2 step mode)
1.000 inch divided by 2000 steps = .0005 axis movement per step.
Use a 2 to 1 timing pulley ratio (like on my Z axis) = .00025 per step.
More than enough accuracy for any of these little mills."
Full step with this setup no pulleys directly running the ballscrews with the steppers would yield a .001 inch per step rate which would make manual g-code entering easier although it may not be as precise as half or quarter step. As far as controllers I would go with Kelings. Lots of features along with protection built into the controller. They are also single axis controllers so if something does happen you only have to buy 1 and your whole machine isnt down because you got the wire pinched in something and smoked the chip in the board. Just my 2 pennys though.