The marchitecture is currently code-named Trinity and denotes a combination of Trinity-enabled motherboard (the 590s) and Trinity-enabled graphics cards. Once that combination is set in motion, the available bandwidth for the cards will go from 4GB/s to 5.2 GB/s, essentially - a legal overclock of the PCIe bus, going from x16 to "x20" or "x22", depending on the final stability tests the company conducted recently.
This isn't the first time Nvidia did a bit of spec tweaking, since the first PCIe products, bridged with the famous BR02 chip - worked at a rate which some reps called "AGP 12X".
We suspect the whole Trinity platform idea was actually a way to get rid of the PCI Express bandwidth bottleneck in multi-GPU configurations. While the details are scarce at the moment, do not expect this feature to be marketeered all over the place.
Just don't be too surprised when all of the sudden, nForce 590 owners with a Trinity-specc'ed board (we suspect all future mainstream and high-end models will support the spec) will get a subtle boost over owners of non-590 based motherboards, such as Intel chipsets or the other Socket AM2 competition. µ
Sign up for INQbot – a weekly roundup of the best from the INQ