Intel 8th Gen CPUs with Vega Graphics

Many had their minds blown when news about Intel and AMD partnering up to make a Intel based chip with onboard Vega graphics spread around the internet. Well today we are getting a full detailed lineup of the new Intel dubbed G series processors. Full specs, release dates, and more.

The Intel Coffee Lake chips and Vega graphics will be interconnected through the EMIB (embedded interconnect bridge), a technology designed by Intel allowing for a PCIe 3.0 8x link between the CPU and GPU. Plenty for modern mid range GPUs and generally even enough for super high end chips.

There will be two variants of the integrated Vega GPUs, one dubbed the Vega M GL and the other dubbed Vega M GH (graphics low, graphics high I presume). Vega M GL will have 20 compute units and be clocked at 931MHz base with a 1011MHz boost clock. While Vega M GH will have 4 more compute units at 24, and a clock speeds of 1063MHz base and 1190MHz boost. Units using the Vega M GL will have a TDP of 65W while Vega M GH will be set at 100W. Both GPUs have access to 4GB of HBM2. Unfortunately the chips with Vega M GL are not just i5 nomenclature, meaning you will have to check which GPU your chip actually has, despite the i7 models and i5 models having the same amount of cores and threads across the lineup. Albeit, the i5 model has 6MB of cache vs the 8MB on the i7 models.

You can view the full lineup in the image blow from Intel.

Interestingly enough the press release also talks about Intel H series processors, mentioning high end discrete graphics which are attached to the CPU much like the G series chips. No specs are official information seems to be given about the H series lineup though, so we will have to wait and see. A large portion of articles seem to be confused as the newest lineup is the G series, as picture above. Although, it’s not totally clear if the chips featuring Vega M GH are the H-series chips, definitely some confusing nomenclature here if you ask me.

All variants of the G series chips will have Intel onboard graphics, meaning you can use the iGPU for lighter tasks getting substantial battery life improvements, and then the discrete Vega M chips for higher end tasks. My question remains, how easy is it to swap between the two graphics chips? We have seen this tech in the past with NVidia Optimus, but it was destroyed with the release of the 10 series as those are all desktop based GPUs, plus it never worked quite right. I’m curious if Intel is going to be able to pull of what NVidia did, but in a more seamless way. Manually choosing which tasks use the dedicated GPU is a big frustrating.

A couple small side notes:

  • The Vega chips still support 6 displays, so in theory you could run 6 displays off a notebook, something NVidia can’t even do with their desktop GeForce GPUs. Not that it’s necessary or even likely that you will need 6 displays, but the option is certainly nice.
  • All benchmarks in the linked below Intel Newsroom PDF were done PRIOR to the Spectre and Meltdown patches. Intel notes this subtly in the PDF, but it’s something important as certain instances can see decent performance drops after the patch.

P.S. Intel please stop using the term “Mega-tasking”, it’s undefined and quite annoying.


Source: Intel Newsroom

Leave a Reply