Intel’s second-generation Arc B580 GPU beats Nvidia’s RTX 4060 for $249

Turnover on the prime of the corporate is not stopping Intel from launching new merchandise: Today the corporate is asserting the primary of its next-generation B-series Intel Arc GPUs, the Arc B580 and Arc B570.

Both are decidedly midrange graphics playing cards that may compete with the likes of Nvidia’s GeForce RTX 4060 and AMD’s RX 7600 collection, however Intel is pricing them competitively: $249 for a B580 with 12GB of RAM and $219 for a B570 with 10GB of RAM. The B580 launches on December 13, whereas the B570 will not be obtainable till January 16.

The two playing cards are Intel’s first devoted GPUs based mostly on its next-generation “Battlemage” structure, a successor to the “Alchemist” structure used within the A-series playing cards. Intel’s Core Ultra 200 laptop computer processors had been its first merchandise to ship with Battlemage, although they used an built-in model with fewer of Intel’s Xe cores and no devoted reminiscence. Both B-series GPUs use silicon manufactured on a 5 nm TSMC course of, an improve from the 6 nm course of used for the A-series; as of this writing, no built-in or devoted Arc GPUs have been manufactured by one in every of Intel’s factories.

Both playing cards use a single 8-pin energy connector, at the least in Intel’s reference design; Intel is providing a first-party limited-edition model of the B580, whereas it seems to be like companions like Asus, ASRock, Gunnir, Maxsun, Onix, and Sparkle can be liable for the B570.

Compared to the unique Arc GPUs, each Battlemage playing cards ought to profit from the work Intel has put into its graphics drivers during the last two years—a mixture of efficiency enhancements plus translation layers for older variations of DirectX have all improved Arc’s efficiency fairly a bit in older video games since late 2022. Hopefully consumers will not want to attend months or years to get good efficiency out of the Battlemage playing cards.

The new playing cards additionally include XeSS 2, the next-generation model of Intel’s upscaling expertise (analogous to DLSS for Nvidia playing cards and FSR for AMD’s). Like DLSS 3 and FSR 3, one in every of XeSS 2’s predominant additions is a frame-generation characteristic that may interpolate extra frames to insert between the frames which might be really being rendered by the graphics card. These sorts of applied sciences are inclined to work greatest when the playing cards are already working at a fairly excessive body price, however after they’re working effectively, they’ll result in smoother-looking gameplay. A associated expertise, Xe Low Latency, goals to cut back the rise in latency that comes with frame-generation applied sciences, just like Nvidia’s Reflex and AMD’s Anti-Lag.