Graphics chip maker Nvidia has been teasing something big for the last few days. After months of leaks and speculation, the new generation of Nvidia cards, the successor to the ultra-dominant GTX 1080 line, is here: the RTX 2080. The new line of cards hits shelves on September 20, but you’d better strap into a bucket seat before I give you the price.
The Nvidia Founders Edition versions of the cards are are follows:
The mid-range RTX 2070 will run you $599 when it releases. This is currently the lowest-end card Nvidia is repping. It’s safe to assume 2060 and 2050 cards are coming later, but they’ll probably be GTX editions instead of RTX – more on that later.
The RTX 2080 starts at $799, and it’s big brother, the RTX 2080 Ti, at $1,199.
There’s some question about these prices. These are the prices on Nvidia’s site as of the time of this writing. But during the keynote, Nvidia Jen-Hsun Huang quoted prices of $499, $699, and $999 respectively. Let’s cross our fingers and hope for those lower prices.
The 2080 cards are already up for pre-order on Nvidia’s site, and third-party manufacturer EVGA already has a few different cards up that start at $749 and $1149, just a hair below Nvidia’s asking price. Other manufacturers are likely soon to follow.
You may have noticed that Nvidia has swapped out the GTX that has preceded its graphics card models for years now with a new scheme: RTX. These cards are the first to feature built-in support for real-time ray tracing. Ray tracing has been something of a holy grail for video games for a long time now. 3D-animated movies like The Incredibles and Toy Story have been using the technique since forever, but it’s an extremely expensive technique that previous cards haven’t been able to handle for even the oldest polygonal games.
Ray tracing is a pretty descriptive name, as it refers to the act of tracing rays of light in a three-dimensional scene. Light hitting a table through a window doesn’t just hit the table, it bounces off the table and illuminates the room based on the material the table is made of, what sort of material is on the walls, and that kind of thing. Right now, game developers simulate that stuff as best they can. They can simulate stuff that’s on-screen, but stuff off-screen has to be “culled” or cut out, to make rendering what’s on-screen possible. So if a light clicks on off-screen, or a gun goes off, that won’t render on your Storm Trooper’s armor or the side of the car you’re driving. Instead, you get the game developer’s best guess about what that reflection would look like. The scenes we see now are two-dimensional simulations of 3D scenes.
Nvidia showed off a bunch of impressive demonstrations of how this all works during its keynote presentation preceding Gamescom 2018. It’s question now of how many games will actually support this new technology out of the box. Nvidia specifically showed off demos for Shadow of the Tomb Raider and Battlefield V, and we’re willing to bet Nvidia is going to be very eager to get other developers to look at the technology in the coming months.
Even without that support, the cards are significantly more powerful – they’re about 25%-33% more powerful in terms of sheer floating-point performance, but Nvidia is careful to point out that with the addition of things like the ray-tracing RT cores and the AI-powered tensor cores make a 1:1 comparison to previous cards misleading.
These cards will be available to put into your system – if you can afford it – starting September 20.
Note regarding the gallery below: These images are not RTX-enhanced.