For $799 and $1,199 (Founders Edition)
At an event on Monday in Cologne, Germany, just before Gamescom kicks off, Nvidia finally announced what we have all been eagerly anticipating—a new generation of consumer graphics cards based on its "monster" Turing architecture, with ray-tracing and deep learning capabilities. Specifically, Nvidia unveiled three new cards, the GeForce RTX 2070, GeForce RTX 2080, and yes, the GeForce RTX 2080 Ti.
This is the first time that Nvidia has announced a "Ti" variant simultaneously with a new architecture. Typically the Ti models come out several months after Nvidia fleshes out its product stack. Not this time though. For this round, gamers can jump straight to the top. The cards are now live on Nvidia's website, and hoo boy, they aren't cheap.
Pricing shakes out as follows:
- GeForce RTX 2080 Ti Founders Edition: $1,199
- GeForce RTX 2080 Ti Reference: $999
- GeForce RTX 2080 Founders Edition: $799
- GeForce RTX 2080 Reference: $699
- GeForce RTX 2070 Founders Edition: $599
- GeForce RTX 2070 Reference: $499
Nvidia is taking preorders for the GeForce RTX 2080 Ti and GeForce RTX 2080 right now, and will ship the cards "on or around" September 20, a month from today. The GeForce RTX 2070 is the only one of the three that's not up for preorder yet—Nvidia says it will be available in October.
Here are the specs for each card:
One thing that's interesting is how Nvidia is approaching its Founders Edition SKUs this time around. FE cards were essentially reference designs with Pascal, but for the Turing launch, the FE models have overclocked boost clocks. That doesn't mean we won't see overclocked models from Nvidia's hardware partners. However, Nvidia's own FE variants, which it sells on its website, kick things up a notch at a $100 premium over reference cards. Nvidia says there are other reasons for the price premium, including an improved cooling setup, but we'll have to wait and see how the retail parts stack up in feature and performance.
On the GeForce RTX 2080 Ti, the FE model's boost clock is 1,635MHz, which is 90Mhz faster than reference. Likewise, Nvidia goosed the boost clocks on its GeForce RTX 2080 FE and GeForce RTX 2070 FE by 90MHz as well, to 1,800MHz and 1,710MHz, respectively.
Beyond the raw specs, Nvidia is heavily emphasizing ray-tracing capabilities from its Turing architecture that powers these new cards. Turing is built for ray-tracing, Nvidia says, and offers 10 times better ray-tracing performance than Pascal. We'll have a deeper dive into all this later, but the general takeaway is that Nvidia's RTX cards are capable of producing better lighting effects than previous generation GPUs.
The Turing architecture also includes Tensor cores that can help with deep learning applications, and Nvidia discussed a new DLSS algorithm that can help produce higher quality upscaling. Again, we'll need to see this in action before we can render a final verdict, but potentially a game could run at 1080p and use DLSS to get near-4k quality without any real performance hit.
Finally, Nvidia CEO Jensen Huang mentioned the ability of Turing to do simultaneous floating-point and integer calculations, at 14 TFLOPS/TIPS each. Graphics cards have to do a lot of address generation when loading textures, and the simultaneous FP + INT calculations could provide a serious benefit to performance, even in games that don't use ray-tracing. Jensen mentioned Turing being "1.5 times faster" than Pascal at one point, seemingly referring to the concurrent FP and INT capability, and if that proves to hold true in most games, even the RTX 2070 could match GTX 1080 Ti performance. Again, that's a an 'if' right now, but Turing is looking potent.
0 nhận xét:
Đăng nhận xét