Intel has unwittingly published details of an unannounced CPU using AMD Radeon RX Vega images integrated directly on the chip package. In November 2017, Intel and AMD sent the PC enthusiast community into a frenzy with an unparalleled joint announcement that the two rivals are collaborating to start one or more such products. At the moment, Intel only stated that these CPUs would slot into its 8th Generation lineup with a launch in the first quarter of 2018, leading to speculation of a launch at the CES trade show in January.
That now seems to be spot on, with a premature listing of the Intel Core i7-8890G appearing on a page comparing the overclocking features of various high-end CPUs on Intel’s India website. It had been spotted by Anandtech and confirmed by Gadgets 360, but has since been removed. According to the listing, the new CPU will include a ‘Radeon RX Vega M GH GPU’ as well as Intel’s own integrated HD Graphics 630 GPU. This confirms that the Radeon GPU is essentially a different GPU that stocks the physiological CPU bundle, but is not a portion of the actual CPU expire in precisely the same way that integrated GPUs are. Intel will probably find a way to balance loads so that the Radeon GPU may be shut down entirely when not needed, to conserve power and reduce heat output.
The supposed Core i7-8809G appears like it will be a quad-core processor with Hyper-Threading for eight powerful threads, using a clock rate of 3.1Ghz. It isn’t known whether that is a foundation clock speed or maximum turbo rate, but it is more likely to be the prior dependent on the comparison to other CPUs. It has an 8MB cache and a 100W TDP. The listing is part of a table which includes all of Intel’s current-gen unlocked desktop CPUs in addition to the 7th Gen X-series lineup, all of which are overclockable. It’s uncertain whether the Core i7-8809G is also a socketed desktop CPU, and if so, whether it will require a new socket and motherboards. Intel was expected to target the high-end notebook section for this launch, but the 100W TDP raises questions of thermal control in a laptop form factor.
The fact that this really is a quad core part, and the Intel HD Graphics 630 name (as opposed to Intel UHD Graphics 630) point to Intel’s Kaby Lake or Kaby Lake Refresh architectures as the basis of the new chip, rather than the newer Coffee Lake. Anandtech references a codename ‘Kaby-G’, which will be further evidence of this CPU’s provenance. No specifications concerning the GPU itself have been disclosed, such as the number of cores or amount of RAM, though it makes sense that AMD will be harnessing its own HBM2 memory technologies. Expected functionality is therefore completely unknown.
For AMD, the GPU is a semi-custom design just like the ones it supplies to the producers of different gaming consoles. We know that Vega on the background computer runs hot and draws a whole lot of power, therefore it will be interesting to determine how the architecture performs on this level.
Shortly after the announcement of this new joint effort between Intel and AMD, Raja Koduri, the former Chief Architect of AMD’s Radeon Technologies Group who oversaw the development of Vega, announced that he had been jumping boat and taking up a new job at Intel. He is currently the Senior Vice President of a newly formed division within Intel known as the Core and Visual Computing Group.