It’s becoming increasingly clear that Nvidia is going to knock one of the park with Ampere. The company is equally certain of this — so much so, that it’s showing the hardware openly in head-to-head match-ups against the RTX 2080 Ti.
In Doom Eternal at 4K, the RTX 3080 rips through the game at frame rates from the 130s to 170+. It’s impressive as hell, and it speaks to just how capable these new video cards are. There are going to be some regretful RTX 2080 Ti owners out there, though I’d argue that while the RTX 2080 Ti is no longer the fastest card money can buy, it’s still an incredibly fast piece of equipment. The GTX 1080 Ti has held its value extremely well since debuting 3.5 years ago. Ampere’s ray tracing capabilities are expected to outperform Turing’s by some margin, but the RTX 2080 Ti will hold value in that category longer than any other Turing GPU.
The only thing I’d note about this comparison is that id’s engines have a reputation for running very well on a wide range of hardware. One side effect of this, however, is that it’s performance rankings can be a bit different than what other engines show. I don’t think it’s unreasonable for the RTX 3080 to be 1.5x faster than the RTX 2080 Ti, on average, but a single point metric absent any context is only a good measurement of itself.
I do not pre-judge hardware and do not recommend pre-ordering anything, hardware or otherwise, but I have to admit that Nvidia has got me curious. Ampere represents a much larger leap ahead for the company on many fronts than the degree of improvement we saw from Maxwell to Pascal, or Pascal to Turing.
Of course, the big question everyone is going to be asking is “Will AMD match it?” Honestly? I don’t know. The gameplay videos we’ve seen from the PlayStation 5 and Xbox Series X have been impressive, but Nvidia’s pricing is extremely aggressive. In fact, I don’t recall any time the company has offered this kind of enormous price/performance leap. The closest I recall, without going back to check reviews, would be if you compared Kepler (first-generation 28nm) with Pascal (first-generation 16nm) to capture the impact of the node shrink. Pascal’s performance improvements over Maxwell were quite good, and the prices on the cards were considered competitive and fair for what they offered, but this jump looks unprecedented.
It was already going to be an interesting holiday season, with two consoles, strong refreshes from Intel and AMD, and two new GPU architectures hitting simultaneously. With Nvidia’s Ampere pricing, we’re looking at a lot of fun launches from here until Christmas.
Now Read: