Intel Kaby Lake + AMD Radeon product *spin-off*

  • Thread starter Deleted member 13524
  • Start date
I kind of behaves like an APU for most intents and purposes, though.
No it doesn't, it behaves exactly like discrete GPU would. The Vega M is connected to the Kaby Lake via standard PCIe x8 bus, only thing you really gain from the MCM is some space on the motherboard and apparently some power benefits from shared power supply to both chips instead of discrete ones for each.
Only MCM in near future that might behave like APU would be the fabled AMD Zeppelin + Greenland (Vega) -HPC-APU-thing, since it's supposed to have coherent 100 GB/s link between the GPU and CPU, but I think it starts to look like that project got buried.
 
No it doesn't, it behaves exactly like discrete GPU would. The Vega M is connected to the Kaby Lake via standard PCIe x8 bus, only thing you really gain from the MCM is some space on the motherboard and apparently some power benefits from shared power supply to both chips instead of discrete ones for each.
Only MCM in near future that might behave like APU would be the fabled AMD Zeppelin + Greenland (Vega) -HPC-APU-thing, since it's supposed to have coherent 100 GB/s link between the GPU and CPU, but I think it starts to look like that project got buried.

What I meant is that from the point of view of an OEM or end-user, it is purchased, assembled, and used much like an APU. The fact that it isn't really one is ultimately pretty transparent, just as all MCMs behave essentially like monolithic dies in most (if not all) cases. Unless you look at the physical package, you won't notice anything.
 
and apparently some power benefits from shared power supply to both chips instead of discrete ones for each.
I don't see how that would be possible, as both GPUs and CPUs vary their core voltage according to load these days. You'd need separate VRMs for that reason alone, plus would mobile vega run exactly same voltage profile as Intel's mobile CPUs anyway?
 
Intel has released a picture of the Kaby Lake G NUC PCB(Look at the I/O on that thing...)
i6omg2K.jpg

For anyone interested you can Pre Order it here
 
Well that looks an awful lot like a Mac Mini does on the back.
multiple mini-displayport/ firewire 2 ports, plus the USB-c / firewire 3 ports.
not sure about the dual ethernet ports though, whats up with that?
 
Intel has released a picture of the Kaby Lake G NUC PCB(Look at the I/O on that thing...)
For anyone interested you can Pre Order it here

Should we draw any conclusions based on the fact that it is shown next to a Mac keyboard? I mean, the part was probably meant for Apple all along. Makes for a great Mac Mini.
 
As explained in the post above, that's the motherboard for the Hades Canyon NUC, not a Mac Mini.

There's no way an apple device would have that many ports. Apple hates physical ports.
And the latest Mac Mini uses 15W CPUs, so unless they're making significantly thicker Mac Minis I don't think they're getting this Kaby Lake G.

KBL-G is probably going into 15" Macbook Pros and iMacs.
 
As explained in the post above, that's the motherboard for the Hades Canyon NUC, not a Mac Mini.

There's no way an apple device would have that many ports. Apple hates physical ports.
And the latest Mac Mini uses 15W CPUs, so unless they're making significantly thicker Mac Minis I don't think they're getting this Kaby Lake G.

KBL-G is probably going into 15" Macbook Pros and iMacs.

I know that is Hades Canyon NUC PCB. I just find it funny that they took the picture right next to an Apple keyboard. Totally unnecessary.

I meant that KBL-G were well suited for the Mac Mini, not that it had anything to do with that PCB.

Going off-topic the current Mac Mini has 11 ports on the back, not including power.
 
Last edited:
Anandtech review of the Hades Canyon NUC

Intel's decision to route all six display outputs to the vastly faster and generally more capable Radeon RX Vega M GPU makes perfect sense for a desktop. But the one area where AMD's latest GPU still trails Intel is in the media decode block. The Vega GPU can't decode VP9 Profile 2 - so no YouTube HDR support - and more importantly it doesn't support the Protected Audio Video Path technology required for UHD Blu-ray playback. The latter unfortunately came as a bit of a surprise to even some at Intel, as the company was claiming as recently as CES 2018 that the Hades Canyon platform would support UHD Blu-ray playback.

That's pretty funny... and sad.
 
It probably will, IIRC AMD said not so long ago that media update with UHD BluRay support is coming soon and Hades Canyon isn't officially for sale yet, is it?
I'm not sure, any reference to that statement? Was it regarding Vega or RR?
 
The latest news were about PlayReady 3.0 becoming available to all GPUs with HDMI 2.0 (Polaris and later).

https://www.anandtech.com/show/12442/amd-plans-playready-3-support-for-polaris-and-vega-gpus-in-2018

I didn't know UHD Blu-Ray required a new PAVP, but why would the Hades Canyon engineers worry about Vega M not supporting UHD Blu-Ray playback?
There's no optical storage in the NUC, and probably neither will there be for any Kaby G solution, as top-end laptops are all dissing optical storage.

Whoever wants UHD Blu-Ray just needs to purchase a dedicated set-top player or a Xbone S. An external USB drive will cost almost as much.


I find that particular review from AT a little bit baffling, since Ganesh manages to avoid comparing the Vega M against the nvidia solutions that Intel themselves claimed that Kaby G was competing against: GTX1060 Max-Q and GTX1050 Ti.
Instead we see comparisons made against dozens of old Maxwell chips and the only current nvidia GPUs in there are the GP104.
 
Maybe they're more concerned with it playing rips from UHD BluRays? Though that all depends on what re-encoding is done with the video/audio payloads.
 
Or the CPU/GPU package being used on a motherboard that is intended for HTPC use with UHD Blurays?
 
Wouldn't they be better off using Intel's GPU for playback anyways? That should be the more power efficient solution for video playback and avoid firing up the dGPU. Vega is efficient, but Intel's solution should still be better in most cases. Vega only makes sense if some app is using SSG style behavior.
 
Wouldn't they be better off using Intel's GPU for playback anyways? That should be the more power efficient solution for video playback and avoid firing up the dGPU. Vega is efficient, but Intel's solution should still be better in most cases. Vega only makes sense if some app is using SSG style behavior.
My impression is that the display outputs are wired to the Vega GPU and you can't simply just use another GPU for protected path output.
 
Back
Top