The Intel Xe graphics project is based on top of over ten years of experience making integrated GPUs and its related software stack. As the company is about to create a complete gaming GPU, it is highly expected that the new Xe HPG graphics card could compete with AMD’s Big Navi and Nvidia Ampere (RTX 30 series), ray tracing, and others.
in the beginning of 2021, Intel started shipping the first generation Xe desktop graphics cards to system builders. These are relatively modest performers that can only compete with some entry-level mobile graphics chipsets for laptops, but they’re a great start and more compelling than something like Nvidia’s resurrected GeForce GT 710.
Meanwhile, Intel engineers have been struggling hard on an upcoming Xe microarchitecture that will scale from integrated graphics on mobile CPUs up to the data center. During Intel’s Architecture Day 2020 event, the company depicted that it is working on Xe HPG chips that will pack GDDR6 memory and support for ray tracing.
Gamers and PC users have been curious for details on its development, but it was only last month that we have got a tweet from Intel’s chief GPU architect, Raja Koduri. He showed that the company is already checking a Xe HPG graphics card using 3DMark’s new Mesh Shader performance test.
Today, a first real hint is here thanks to the PC world – Intel’s new gamer-focused GPU must hit at least GeForce RTX 3070 levels of performance as per synthetic workloads.
From 2012 to 2021 – same Intel Folsom lab, many of the same engineers with more grey hair , I was at Apple back then, getting hands on with pre-production crystalwell, 9 years later playing with a GPU that’s >20x faster! pic.twitter.com/RgmRJuhOXw
— Raja Koduri (@Rajaontheedge) March 12, 2021
Koduri shared an image depicting many people that worked on Intel’s Crystal Well tech years back in 2012, now working on Xe HPG at the same Intel Folsom lab, and seeing 20x the performance in synthetic benchmarks. Crystal Well refers to the Intel Iris Pro 5200 integrated graphics that came with an embedded DRAM cache and can potentially obtain points from 1,350 to 1,450 points on the 3DMark Fire Strike test.
Nvidia’s almost impossible to get RTX 3070 and RTX 3080 graphics cards can get about 25,000 points and 30,000 points respectively. Intel’s own Xe Max can get about 5,800 points.
It is not explaining the level a Xe HPG graphics card would compete with Nvidia’s offering when it comes to gaming performance. Still, there are a few things to reinforce the validity of this comparison. The first is that Intel will utilize a different process node as compared to its 10nm SuperFin making it a potential buyer of TSMC’s excess wafer capacity.
The second important thing is that Intel’s Xe HP architecture is designed to scale well using “titles” that each owns 512 integration units or about the equivalent of 4,096 shader cores for Nvidia and AMD graphics solutions. The Xe HPG will certainly cut various surplus features of its server counterpart to support more integration units operating at high clocks. Moreover, Intel could use 2 HPG tiles associated with an embedded multi-die interconnect bridge (EMIB) to provide beast level power, just like SLI and Crossfire but with improved scaling and less halted.
In any of the cases, Intel could be working hard to hit AMD and Nvidia with the best graphics solution at a time when both are trying hard to meet the demand for their new provisions. If the performance, power consumption, and pricing are right, Intel’s Xe HPG graphics card might be used in many gamer rigs whenever it comes.