During CES earlier this month, Intel took the wraps off of its 8th Generation Core Processor with Radeon RX Vega M Graphics product, an innovative processor that combines the power of a quad-core Intel CPU and a discrete GPU from AMD on a single package. The technology behind the product is nearly as impressive as the business decision to make it happen, but unfortunately it looks like Intel and its partners might be misusing the potential that the processor codenamed Kaby Lake-G provides.
Under the hood, the Kaby Lake-G processor is a marvel. By combining the leading performance processor in the mobile space with a discrete-level graphics solution on the same package, Intel is able to build a part unrivaled in the consumer space. The graphics chip comes from AMD and is a semi-custom design based on the Vega architecture. It is physically connected to the processor via 8 lanes of PCI Express through the silicon interposer. The GPU has a single stack of 8GB of HBM2 memory (the same used on AMD’s latest Radeon RX Vega series of graphics cards) connected through Intel’s new EMIB (embedded multi-die interconnect bridge). This EMIB allows for a high bandwidth interface between the two chips (required for high-bandwidth memory) while allow for low z-height package design (critical for mobile form factors aiming to thinner).
Seeing Intel purchase a part directly from AMD in the form of a semi-custom graphics chip to integrate on its own processor design, in addition to depending on AMD to provide the necessary driver support for Intel to repackage and distribute, is still a shocking revelation. This gives Intel its first truly high-performance processor design with CPU and GPU strengths, though it does indicate where Intel believes it stands in today’s market in regards to graphics performance.
But what has surprised me the most about this new processor is Intel’s decision to not market this product to gamers.
Kaby Lake-G will be available in two flavors: a 100-watt version and a 65-watt version. Claims from Intel put the 100-watt version on par with the graphics performance of an NVIDIA GeForce GTX 1060 and the 65-watt version at a GTX 1050. These are impressive claims and should give any notebook vendor an incredible opportunity to build a system with gaming-class performance but without the need for two dedicated, discrete chips. This should allow for smaller PCB design, larger batteries, thinner form factors, etc.
At CES, Dell and HP showed new notebooks integrating the Kaby Lake-G processor. The Dell XPS 15 2-in-1 and the HP Spectre x360 15 are both impressive systems in their own right, combining solid design with thin form factors and great feature sets. However, when I talked with both parties about the planned market for the machines, they had the same answer: content creation users. When I asked about gamers, both companies seemed disinterested in the idea, saying these products instead “could be used for gaming.”
To be clear, these designs and the Intel Kaby Lake-G processor will indeed be excellent for content creation, especially any users that take advantage of software that can utilize the GPU. Adobe Premiere, rendering tools, CAD, etc. will all see improvements thanks to the addition of the discrete-level graphics on the new part. But there are two big reasons why Intel and its partners should be leaning more heavily into the gaming angle.
First, the timing is perfect. With the increase in prices for gaming graphics cards, and system memory, causing many DIY users to put off or avoid a new system purchase or upgrade thanks to the shortages caused by cryptocurrency mining, any alternative solution today will get a lot of attention. I have already explored how the current market is pushing consumers toward OEM PCs and gaming notebooks, and a sleek, compact, and innovative design based on Kaby Lake-G would be an excellent choice. Gamers that have been beaten up on pricing because of coin mining will be looking at pre-built machines, and gaming notebooks, more than any time in the history of gaming notebooks. Intel’s latest processor combination can be implemented in capable systems that are as gorgeous as they are performant.
Our second reason centers on what has become Intel’s enemy #1: NVIDIA. There is no denying the envy that exists from Intel when looking at the rise in value and technological advantage that NVIDIA has created in the fields of machine learning, artificial intelligence, autonomous driving, and event high-performance compute. While Intel is on track to bring competitive options in those spaces in the coming years, it could be using the Kaby Lake-G part to compete with NVIDIA on its home turf: discrete graphics chip for gaming. NVIDIA essentially owns the mobile graphics market and (nearly) any machine currently shipping that is targeting gaming will include a discrete GeForce part in it. The Kaby Lake-G solution will be able to offer competitive performance with the largest area of the mobile gaming market (GTX 1050-1060 performance level) and Intel can leverage its combined package and design as an advantage. Not only that, but some good old price negotiations couldn’t hurt either.
And from my perspective, the design of Kaby Lake-G just makes it good at gaming. Intel has discussed openly its ability to balance CPU and GPU power draw to maximize performance efficiency for games, how it plans to properly support gamers with day-zero driver support for new titles, and quotes gaming performance as its flagship metric to the press. But the first wave of products based on the processor don’t seem to have even considered that angle.
This is why Intel’s decision to not focus on the gaming angle for the 8th Generation Core Processor with Radeon RX Vega M Graphics is odd and out of place. Yes, there are multiple uses for this new processor family but I think Intel, Dell, and HP are missing out on a crucial opportunity with gamers and the eager gaming environment .