Intel could make waves in the GPU scene with its integrated, or add-in, Xe graphics like how AMD rocked the CPU world. The US-based chipmaker revealed the roadmap to the new Xe GPU architecture on its latest ‘Architecture Day’.
Xe graphics are Intel’s 12th generation of integrated graphics that follows the Gen11 ‘Iris Plus’ graphics found in Ice Lake CPU devices. The Xe lineup will debut on devices, alongside Tiger Lake SoCs (system on a chip), as soon as September.
On Architecture Day, Intel also put rest to speculations regarding the rumored gaming-focused Xe graphics. Their announcement confirms that the Xe family will consist of four variants. Find Intel’s GPU Architecture Strategy below.
- Xe-LP: The lowest tier of entry-level integrated graphics for mobile devices like laptops. The ‘Low Performance’ lowers battery consumption while still providing performance boosts over previous Intel integrated graphics.
- Xe-HPG: The latest announcement to the Xe graphics lineup where the ‘G’ stands for gaming. More on the Xe-HPG later.
- Xe-HP: While the HPG focuses on gaming and graphics performance, the HP (High Performance) targets Data Centers and AI. These high-end chips have features like FP64 and multi-tie scalability.
- Xe-HPC: The highest-end exascale Xe GPU will power supercomputers. Codenamed ‘Ponte Vecchio,’ the HPC (High-Performance Computing), will use a more ‘enhanced’ process than the HP.
Intel Xe graphics
After falling behind to competitors, Intel’s Xe graphics are finally going to use a 10nm SuperFin process in the market, AMD, has already moved on the 7nm process but whatever.
The SuperFIn technology combines Intel’s refined FinFET process with a “super metal insulator metal capacitor.” The technology will allow Intel to deliver significant speed boosts ever within a single node.
The Gen12 Xe GPUs will deliver at least twice the performance of Gen11 Iris Plus graphics. One area that the Xe graphics have evident upgrades over is Execution Units. The Xe integrated graphics will have 96EUs over the top-end G7 Iris Plus.
The definite performance boosts will mean notebooks with Tiger Lake processors (integrated Xe graphics) can smoothly run games like Battlefield V at HD settings.
Perks of the job! Took a prototype Tiger Lake system for a spin on Battlefield V to stretch its legs. Impressive thin and light gaming perf with Xe graphics! Early drivers/sw, but it’s the first time I’ve seen this game run like this on integrated gfx. More later this year! pic.twitter.com/f1Qlz2jMyB
— Ryan Shrout (@ryanshrout) June 17, 2020
Expect enthusiasts to put the Tiger Lake devices through gaming tests come September.
Why will the Xe-HPG give AMD and Nvidia a hard time?
As mentioned previously, Intel announced the Xe-HPG on its latest Architecture day. The dedicated GPU will have support for GDDR6 memory, scalability, and, get this, ray-tracing.
Only Nvidia’s GeForce RTX cards have support for ray-tracing, and AMD has hiccuped to bring the feature to portable devices. AMD will be bringing support for ray-tracing on the Sony PS5 and Xbox Series X, through Big Navi GPUs.
Additionally, the Xe-HPG will consist of all technologies found in the Xe-LP, Xe-HP, and Xe-HPC.
- The add-in graphics card will have the low-consumption foundations of the Xe-LP,
- Through Xe-HP tech, the Xe-HPG will be able to scale to hundreds of GPU execution units. Expectedly, there will be no support for non-consumer features, like FP64 and multi-tiling, found in the Xe-HP/HPC.
- Furthermore, “compute efficiency” from the Xe-HPC will allow the Xe-HPG to maximize overall clock speeds.
Ray-tracing is currently a high-end GPU feature, which will become a standard in the years to come. Intel’s Xe-HPG supporting the functionality already puts it in the range of Nvidia’s and AMD’s best gaming GPUs.
In the end, Intel’s Xe graphics will bring more competition to the graphics card space. As seen with AMD’s competitiveness in the CPU space, it is a win-win situation for customers.