Intel getting ready to face nVidia and AMD
Its Arc brand will include graphics processors for desktops and laptops - tentative roadmap also revealed
It's no secret that the nVidia/AMD duopoly could use some competition in the consumer space: for far too long all we've been doing is watching these two companies trying to outdo one another in specific price points while the cost of graphics cards themselves has been steadily climbing up (even before the COVID-19 pandemic made those products so hard to find). It's not easy for any one company to face AMD or nVidia with a competing product, of course, as they have both invested unthinkable resources in GPU design for more than 25 or 30 years now. But if there's one company capable of doing that, it's Intel - and, by the look of things, we'll get to see just how it intends to do that in just a few months' time.
Intel just announced its "Arc" brand name: it encompasses both the hardware and software side of the company's discrete graphics cards and it’s meant to span multiple generations. The first of the Arc GPUs - the one previously known as DG2 - is codenamed "Alchemist" and is expected to arrive in Q1 2022. Other graphics chipsets, codenamed "Battlemage", "Celestial" and "Druid", will power future generations of Arc GPUs but Intel did not share more details regarding those (it will do that later in 2021).
The Arc GPU range is confirmed to support a number of modern graphics technologies such as mesh shading, variable-rate shading, video upscaling and real-time raytracing. Intel also promises to offer its own AI-accelerated super sampling function, unnamed as of yet, which will function like nVidia's DLSS and AMD's SFR. This is extremely important: it allows PC gamers to play demanding titles at smoother framerates by upscaling each frame from a lower resolution with minimal compromises in image quality. Needless to say, it proved to be one of the most important features of the current generation of graphics cards, maybe one of the greatest ever.
It's not yet known whether Intel's upscaling solution is based on a basic technique, such as AMD's spatial FSR, or a more advanced one, such as nVidia's temporal DLSS. As a matter of fact, nothing about the tech specs of the first Arc GPU was mentioned by Intel, such as processing cores, bus interface, RAM size, clock frequencies, power consumption or output ports. Rumors are making the rounds for months now, placing the top "Alchemist" GPU model (there will be a number of cut-down variants) right up there with the most powerful graphics cards nVidia and AMD offer today - but, then again, in Q1 2022 both of them will be ready to unveil their next offerings.
What we are all eager to find out, of course, is (a) the level of gaming and application performance the Arc graphics cards will be able to hit as first-generation GPU products and (b) how aggressive Intel means to be in terms of pricing. Even if these products don't quite match the ones nVidia and AMD are planning for 2022 performance-wise, offered at a low enough cost they can achieve something equally important: disrupt a product category that has not undergone a true shake-up for many, many years.
Intel may be the only company with the marketing muscle, the client list and the manufacturing capacity to pull this off. The timing will be far from perfect - nobody is expecting shortages in microchips, especially GPUs, to not still be a thing in Q1 2022 - but, then again, market conditions will never be perfect for a bold move like this one. Intel might as well proceed, do its best, gather feedback, make adjustments and prepare for the next step. Since the company seems to be in this for the long haul, it sounds like a plan!