Forget the PS5 and ray tracing – Intel’s GPU roadmap looks e-Xe-mplary
The world of gaming has been in a rapid state of flux over the last couple of years, with key companies including Nvidia, AMD, Sony, Google and Microsoft all announcing new products and technologies at a near frenzied pace.
This year that’s been marked by the growth of ray tracing graphics tech, which first debuted on Nvidia’s RTX line of graphics cards. It’s then set to expand next year with the arrival of the PS5 and Xbox One games consoles, which are both confirmed to feature next generation, ray tracing-ready, AMD graphics.
This week we also saw the arrival of Google’s curious new Stadia game streaming service, which, while a little undercooked on actual games, is definitely an impressive technological achievement.
But for me, while the idea of a new console and more realistic graphics are cool, there’s one not so talked about development that could be an even bigger deal for gamers: Intel’s spiffy new Xe line of GPUs.
Intel Xe hasn’t hit mainstream headlines, but we’ve actually know about it for quite a while. The company’s been pulling an AMD-Zen – and by that we mean quietly discussing its plans without showing a product – with developers at various events for ages.
And while most of the conversation has been about server-side computing, which Intel already has a fairly big footprint in, providing the chips for key services like Stadia – this week the firm revealed key details proving Xe could be a huge deal for mainstream gamers and consumers.
The news broke at the HPC Developer Conference earlier this week, where Intel SVP, chief architect and general manager of architecture, Raja Koduri revealed the company’s initial Xe three tier roadmap. This splits the Xe cards into the below lines:
- Xe-LP: The entry level Integrated series that’s optimized for 5W-20W power consumption
- Xe-HP: The midrange enthusiast dGPU line that’ll take on AMD and Nvidia cards
- Xe-HPC: The top end line for top-end server/stack computational work (non-consumer)
Those descriptions may not sound that exciting, especially given that we won’t see any consumer grade Intel Xe GPUs hit the market until after well after 2021, but it’s still some of the most exciting news to hit the headlines this year.
This is mainly down to Intel’s strategy. If you watched the keynote you’d have heard Kodur talk a lot about Xe’s Multi-Chip Module (MCM) arrangement, nanometer counts and Intel’s OneAPI strategy at great length and in absurd amounts of technical detail. But if you read between the engineering speak there was a very clear message: we want Xe to be super easy to scale, develop and support.
These are three things that could be great for gamers and game developers. They also tie directly into the goals Koduri laid out, during a briefing attended by Trusted Reviews, about Xe’s place in gaming earlier this year.
Related: Best graphics card 2019
There Koduri claimed Integrated Xe graphics would power triple-A games with ease, which is cool. But if the scalable design delivers the way he claimed at HPC, it could offer third party companies much greater freedom than they currently have with Nvidia and AMD architectures. This in theory could let them create a more diverse portfolio of cards for bespoke user cases, like laptops, micro-ITX builds and gaming.
But more importantly it could potentially let them make custom build cards, where the customer picks the specific specs – which would be just as revolutionary as when Michael Dell started doing that with laptops back in the 80s. This is why Xe could be one of the biggest announcements to arrive this year.