The variable nozzle requires a variable cycle engine concept. In other words, it costs in test and software programming time. Before the compressor stator mapping couple be a function of outside temperature (well... density too), fuel flow, engine RPM, and a few 'tweaking variables' that were relatively easy to map.
Now with the variable nozzle, the engine must adadpt to the fan nozzle as the fan nozzle is adapting to either maximize thrust for a given fuel burn. In other words, it requires a FADEC that has significantly more computational power with alternate mappings for the compressor stators in case the nozzle fails full open, full closed, or anywhere in between.
In 2000, the best FADECs were single core PowerPC 630s. Not exactly a barn burning chip. I forget the clock rates, but they were heavily reduced in oder to survive being glued to the motherboard (worse thermal conductivity) and the variable thermal environment of a jet engine.
Also recall that for safety critical flight control (which the engine should be part of), the CPU is usually loaded to 25% of capability as one is required to stay under 50% cycle loading.
Then the software testing....
The I/O data, I take from this link, is 25% to 30% greater with the PW1000G and doing quite a bit more decision making with that software. I can only begin to imagine the size of the FADEC team for the PW1000G and all the variants: