Speaking from personal experience (I used to run both small and big 16Gs on the same car with the same exact setup a while back), the difference between those two turbos in question is really like splitting hair. Although the small 16G "might" possibly spool a bit faster, the spool-up time is affected by so many things (such as level of hacking the MAS, porting the turbo housing and exhaust manifold, compression of the engine, clipping the turbine wheel, A/F ratio and choice of O2 housing and the entire exhaust) that it's extremely - if not impossible - to compare the spool-up time of those two turbos in fair manner. Even if the small 16G may spool a bit faster than the big 16G on the same exact application, the difference won't certainly be in 100s of RPM but most likely only in one 100 RPM range. The top end pull and ability of the turbos to sustain a high level of boost is going to be affected in similar manner on two different applications as well.
When going from the small 16G to the big 16G without changing a thing, the first two gears and spool-up time seemed to be not just nearly, but 100% identical. The only slight difference I could really notice was the bit better top end pull of the big 16G in 3rd gear above 6K RPM at 20+ boost levels. The only real advantage of the big 16G seems to be its slightly better efficiency - which might possibly make some difference when combined with the stock SMIC (that will heatsoak during 3 hard runs anyway) but will most likely disapear with the use of FMIC. As a matter of fact, due to the high efficiency of today's FMICs, some slightly different efficiency levels of the turbos don't play any really cruscial role any longer. Sure, even so, less efficient turbo will produce more hot air when pushed out of its efficiency range - as we all have the tendency to do
with our 14Bs and 16Gs. If one really considers also the huge differences in tuning, thermal and volumetric efficiencies of different setups and engines themselves, then comparing the small and big 16Gs is really pointless - since, as with every turbo, the final results will depend on way too many variables. We all have seen people with 20Gs and Greens barely breaking into 13's and - on the other hand - people with 14Bs knocking on the door of 11's, or people with 16Gs actually running high 11's.
In my opinion, 16G makes a great turbo for autocross and possibly also a rally car, but one seeking a true power with decent spool-up for lights-to-lights street confrontations, 20G is the way to go. Personally, I wouldn't even bother with anything in between - such as the 18G or 17 or 19C. If someone can't choose from three basic levels:
1) A bit more than stock (16G, T28) - upto 360HP,
2) As much as the stock block internals can handle (20G, Green, smaller T3/T4) - 400-450HP and
3) F*ck'em up (Red, L3R, larger T3/T4 or T-series) - 450HP+,
then he's in serious trouble... Just my 0.02.
