The Intel-AMD Core i7-8809G will be overclockable with a 100W target TDP
Ever since Intel and AMD announced their collaboration to create a multi-chip module (MCM), expectations have been sky high as to what would be the specifications of the chip. Now, Intel India has inadvertently leaked some information about what to expect from the new chip via a listing of overclockable processors that includes the new Intel Core i7-8809G.
The 8809G is being listed as a 3.1 GHz CPU with 8 MB L3 cache and a 100W 'target package' TDP. It supports dual channel DDR4-2400 RAM and includes both the integrated Intel HD Graphics 630 as well as the new Radeon RX Vega M GH GPU. There's quite a bit of information in there so let's check it out a bit further.
First up, this belongs to the 'Kaby Lake' series and not 'Coffee Lake' like most new Intel chips, considering this is a quad core CPU. 'Coffee Lake' also has quad core CPUs, but those are 35W chips. Also, 'Coffee Lake' CPUs have the iGPU labeled as UHD Graphics instead of HD Graphics. The other interesting tidbit is about the Radeon RX Vega M GH GPU. While M could imply mobile, we are not quite sure what GH refers to. While we have seen a few leaked benchmarks that pit the Vega GPU in the MCM against the likes of the NVIDIA GeForce GTX 1050, we still do not know the specifications of the chip such as the number of compute units (CUs) or the amount of HBM2 memory it will have. However, one thing that can be confirmed now is that the MCM houses a Vega GPU and not Polaris as was speculated earlier.
Interestingly, the 8809G is being pegged at a 100W 'target package' TDP. Target package is the keyword here as it means that it is a combined TDP for both the CPU and the GPU. Intel, in its original announcement of the MCM, advocated this new chip for use in ultrabooks and everyone knows that ultrabooks and high TDP don't rhyme. This could mean that here are low TDP variants for notebook use in the pipeline as well. Let's look at some math here — since it's a 'Kaby Lake' quad core, consider the CPU TDP to be somewhere in the 45W range similar to the HQ-series we have today. That leaves us with about 55W for the GPU and HBM2 combined. For comparison, the Ryzen 7 2700U has about 10 Vega CUs in a 15W envelope so one can safely assume that the Radeon Vega M GPU will have anywhere between 20-24 CUs.
Finally, the 8809G is being listed among the crop of overclockable CPUs in the table and considering the large target package TDP, it could be a desktop variant. The table does not list the turbo frequency of the CPU. The MCM could also afford a new motherboard package as existing sockets don't cut it for the new design or it could be embedded on the motherboard itself similar to the leaked photos we saw a few months ago.
The news comes an interesting time just when CES 2018 is around the corner. We can presume that we will be seeing more leaks in the run up to the big event so stay tuned for all that and more.