Intel’s integrated graphics technology has come a long way in recent years, making it possible to power multiple 4K displays or play some less demanding 3D games on a computer featuring just an Intel processor with an Intel UHD integrated GPU.
But rivals AMD and NVIDIA haven’t exactly been sitting still — NVIDIA’s entry-level MX130 and MX150 graphics cards for laptops offer a significant graphics boost without adding a lot to the price tag (or taking too big a toll on battery life). And AMD’s Ryzen processors feature integrated Radeon graphics that can run circles around Intel’s Gen9 integrated graphics.
So in the coming year Intel plans to skip over Gen10 altogether and crank things up to 11. The company says its Gen11 integrated GPU will offer up to twice the performance of today’s Intel UHD integrated graphics technology.
Gen11 GPUs will have 64 execution units (up from 24), with support for over a trillion floating point operations per second (TFLOPS).
That should be more than enough to handle 8K video content, and it should let more games run smoothly without the need for discrete graphics. There’s also support for Intel Adaptive Sync smoothing technology for smoother frame rates while gaming.
In other words, Intel’s Gen11 GPU should be roughly on par with the Radeon Vega 8 GPU featured in AMD’s Ryzen 5 2500U processor.
We’ll likely see Gen11 graphics make their debut with Intel’s upcoming 10nm chips based on Sunny Cove architecture in the second half of 2019.
Find more technical details at AnandTech
I am sure the 64 execution units will be limited to the high end processors, like 35w and 65w quad cores. Expect the celeron and pentium to have 8-12 execution units
Speaking of Tower PCs, I doubt it. Anyone who has $500 for i7 could spent $50 for GT1030. Speaking of Celeron users, they need it much more, because they buy cheapest solution and don’t have spare dollars to spend for dedicated GPU.
And for Office use, it’s not important how much cores iGPU has, more important it could play 4k videos encoded with HEVC/H265/10 bit or more/or-any-other-latest-fashioned-videocodec-up-here.
Meanwhile, the Ryzen 7 2700U which is hamstrung by GloFo’s 14nm process gets 1.7 GFLOPS. Putting two and two together, it is logical to assume that next-gen Zen 2 APUs with Navi on TSMC 7nm will be in the 2-4 GFLOPS range. This was just a long overdue upgrade for Intel’s current aging iGPU architecture. So Intel still is fashionably behind the curve with their iGPUs, again failing to aim ahead of the moving target.
I was about to say “who cares”, but then I realised a great a number of devices; MacBooks, MacBook Pros, Tablets, Ultrabooks, Business laptops, Office PCs, and Intel NUC’s actually depend on the performance of these iGPU’s.
You cannot add a dGPU to them, and an eGPU solution seems to be the last thing Microsoft, Intel, and the OEMs care about in 2016-2019. Maybe in 2020 the PS5/Xbox V might help stir this up?
Brad Sams’ at Thurrott.com recently revealed a credible rumor that the next Surface Laptop will feature an AMD processor which further cements the underpinnings of a fundamental shift. A stagnant Intel leaves a market ripe for change across all product form factors.
Add the fact that the only meaningful IPC improvements in Intel’s next CPU architecture are just encryption acceleration and this leaves little confidence about their ability to compete with AMD for the next few years. Meanwhile, Zen 2 also offers double the AVX performance per clock cycle compared to Zen and Zen+ as well as general integer and floating point performance gains in the area of IPC efficiency, making for an Athlon 64-like slamdunk.
Your first instinct was correct. For the vast majority of non-gaming PC-based solutions, iGPU performance is already “good enough.” and the new generation of consoles isn’t going to change that significantly.
It doesnt work that way. From 14 nm to 7 nm is only a 45 percent performance increase, not 2 times. It should also be known that ryzen apus cannot reach full speed at a 15 watt tdp
Comments are closed.