|Specs at a glance: Nvidia GeForce RTX 2080 Ti Founders Edition|
|MEMORY BUS WIDTH||352 bits|
|MEMORY SIZE||11GB GDDR6|
|Outputs||3x DisplayPort 1.4, 1x HDMI 2.0b, 1x USB Type-C (VirtualLink VR)|
|Release date||September 20, 2018|
|PRICE||Founders Edition (as reviewed): $1,199. Partner cards priced at: $1,169.|
Like any piece of expensive technology, a top-of-the-line graphics card comes with all manner of lingo and abbreviation. You’ll need a glossary to wade through the stuff inside (processors, CUDA cores, ROPs), the speeds measured (memory bandwidth, boost clocks, TeraFLOPS), and the results you want from a good card (anti-aliasing, frame rates, higher resolutions).
The two biggest companies in the space, Nvidia and AMD, take different approaches to various GPU elements (and those approaches further differ in the realms of lower-cost and heat-efficient GPUs). But at the very top of the heap, Nvidia rules. And not by a slim margin.
This far into 2018, the company’s last wave of PC graphics cards still leads the market in sheer gaming performance, including March 2017’s GTX 1080 Ti . This $700-ish card remains the 4K graphics solution in the consumer sector, and AMD has yet to touch it in terms of performance. In fact, its late-2017 RX Vega 64 (roughly $550) doesn’t even definitively top the older GTX 1080 (non-Ti, roughly $490).
That state of affairs put Nvidia in an interesting position in August when the company announced a new wave of graphics cards with even more terminology and lingo attached. We weren’t shocked to see one of Nvidia’s favorite topics, ray tracing, emphasized at length. But after the dust and announcements settled, one of Nvidia’s choices of phrase stuck in our mind: “RTX-OPs.”
You read that correctly. Nvidia invented a new form of graphics measurement (and one that piggybacks over the more conventional “FLOPS,” aka “floating-point operations per second”). What’s more, the company’s measurement assumes a perfect 3D-acceleration pipeline: traditional rasterization combined with an intentional tie to multiple Nvidia-only APIs. (Conveniently enough, Nvidia’s new RTX cards feature three processor cores that combine forces to do just that.) Put all of those together, and , Nvidia swears you’ll get a ton of RTX-OPs.
|RTX 2080 Ti||RTX 2080||GTX Titan XP||GTX 1080 Ti||GTX 1080||GTX 1070||GTX 1060||GTX 980 Ti|
|Memory Bus Width||352-bit||256-bit||384-bit||352-bit||256-bit||256-bit||192-bit||384-bit|
|Memory Size||11GB GDDR6||8GB GDDR6||12GB GDDR5X||11GB GDDR5X||8GB GDDR5X||8GB GDDR5||6GB GDDR5||6GB GDDR5|
On its face, that’s pretty rich. So what if Nvidia invented its own measurement standard and declared itself the winner? That’s as if McDonald’s measured the flavor and value of fast food with a term like “Grimaces.” Methinks Wendy’s would lose.
But it’s not an entirely crazy move on Nvidia’s part. Without AMD apparently nipping at its high-end PC heels, Nvidia enjoys some bandwidth (pun intended) to try some crazy things. In this case, those come in the form of proprietary standards for anti-aliasing, ray tracing, and rasterization.
For the most part, none of those new Nvidia standards relies on the traditional “more speed, more CUDA cores” formula of old. Some impressive tech is clearly going into Nvidia’s efforts, including a shift to a 7nm process—and thus making room for new discrete processor cores on the RTX series’ chips. Still, Nvidia is rolling the GPU dice with an RTX-fueled pivot. We’re not just going to be faster than AMD, the RTX series trumpets. We’re going to get our hooks into PC-gaming developers through lower-cost special effects that just happen to be exclusive to Nvidia cards.
The catch, of course, is the future tense.
In reviewing “Founders Editions” of the RTX 2080 ($799) and RTX 2080 Ti ($1,199), I ran the usual gamut of benchmarks, and I did the usual double-checks of various games. But all the while, I turned Nvidia’s sales pitch over and over in my mind—because that’s more or less all I could with the sales pitch.