NVIDIA GeForce RTX Turing Architecture - Inspecting The 2080 & 2080 Ti
GeForce RTX 2080 and RTX 2080 Ti look virtually identical, save for their individual brandings which are located smack-dab in the middle of their fan shrouds. The cards feature heavy-duty, brushed metal all-around, with blackened mounting brackets, heavy-duty etched backplates, and lighted GeForce RTX logos running along the top. Chrome bolts and accents tie it all together, and result in some gorgeous graphics cards in our opinion. We should mention that the GeForce RTX 2070, which is coming a little later in October, has a similar design language, but the cards are slightly shorter overall.
There are dual, 13-blade axial fans on the cards that sit atop a high-capacity vapor chamber (double the size of Pascal's cooling solution), and coupled to a dense array of heatsink fins by a thick baseplate with embedded heat pipes. The entire assembly is rigid and quite heavy and covers the entire front of the PCB. NVIDIA claims the new design can be up to 5x quieter than previous-gen cards, while also offering significantly better cooling performance throughout the voltage / frequency curve.
The GeForce RTX 2080’s base GPU frequency is 1515MHz. The standard edition card has a boost clock of 1710MHz and the Founder’s Edition 1800MHz. The GeForce GTX 2080 Ti’s clocks are somewhat lower, however. The GeForce GTX 2080 Ti’s base clock is 1350MHz with a standard edition boost of 1545 MHz; the Founder’s Edition will boost up to 1635MHz. At those speeds, the GeForce RTX 2080 has a TDP of up to 225W and the RTX 2080 Ti of up to 260W, not including an additional 35W that comes into play when the new VirtualLink / USB Type-C connector is used.
To supply power to the boards, the RTX 2080 packs both 6 and 8-pin PCIe power feeds and the RTX 2080 Ti sports a pair of 8-pin feeds. NVIDIA has also re-vamped the all-digital VRM designs on the cards for smoother delivery, reduce noise, and generally be more robust. The GeForce GTX 2080 has an 8-phase design, while the 2080 Ti is outfitted with 13-phases. Many of NVIDIA’s partners are outfitting their custom boards with even more.
Outputs on the cards consists of a trio of DisplayPorts (v1.4 with DSC 1.2), an HDMI 2.0b port (with HDCP 2.2), and a proprietary VirtualLink USB Type-C port, which offers a 4 lane HBR3 DisplayPort and USB 3.1 Gen 2 over a single connector.
The traditional SLI connectors normally located along the top-edge of high-end GeForce cards are now gone, however, in favor of a new NVLink connector that offers significantly more bandwidth, but we'll cover that a little later.