There seems to be a lot of fuss over boosted 928 dyno sheet comparisons in the last 2 days. I would strongly caution people to make sure you ask exactly how the dyno tests were run before you jump to conclusions on the results.
Standard protocol is to lock the car into one gear as close to 1:1 ratio as possible and then do the pull while remaining in that gear...this takes the gear multiplication out of the picture totally and you are left with a rear wheel number (with driveline frictional loss), but not including the multiplication effect of the gear ratios...or at least it is kept to a minimum.
It's a well known fact that twin-screw (positive-displacement) superchargers trump the torque of any other form of forced induction system off idle and down low, say to about 3000 RPM...(thus the reason they are nice for off-idle and stoplight to stoplight bursts). This is the point at which a turbocharged car typically will start to out-torque a twin-screw car (assuming you fed the hampster(s) properly that day
). Centrifugals rarely if ever match the off idle torque of the twin-screws (remember centrifugals build boost exponentially...whereas the screw super hammers the boost on hard as soon as the throttle is cracked...usually the boost is more of a plateau than a curve on a TS pump). Turbos roll on boost hard at say 3000 rpm and then hold that boost flat until redline. Centrifugal boost curve is exponential...if you are set for 10 psig at 6K rpm you will have roughly the square-root of that psig at 1/2 the rpm....in this case 3psi. At that point the twin screw pump will for sure be on "full song" (10 psig in this example) and the turbos will be ramping on pretty solidly assuming they were sized properly for the application and the car is loaded up.
Consider that the three forms of forced induction all have boost curves which are very different...all other factors equal and neglecting the energy required to turn the various forms of compressors (turbo pulls the least out of the motor to drive itself), the torque curve shape will follow the shape of the boost curve. It's a function of BMEP inside the combustion chamber, pure and simple. Torque multiplied by speed gives power...you can't have HP without torque AND shaft speed. By definition power is a rate of doing work and for shaft power one has to consider the torque and speed to arrive at the resultant power.
Again, I caution people when making these comparisons...you have to know how the tests were run before you make assumptions and comparisons. The protocol used for my pulls are listed in this thread and are "standard practice" for the industry.
Never assume and always ask questions.