Intel Core Processor Combines CPU with Discrete Graphics & HBM2 From AMD

Published by

Click here to post a comment for Intel Core Processor Combines CPU with Discrete Graphics & HBM2 From AMD on our message forum
https://forums.guru3d.com/data/avatars/m/156/156133.jpg
Moderator
This is kind of exciting.
https://forums.guru3d.com/data/avatars/m/63/63170.jpg
The whole world and his dog are stating "Hell hath Frozen over...." I'm excited to see what type of GPU they are putting in this. how powerful it is, and how many shaders, etc... not sure about the power consumption tho. it could be 45W all told, or 45W for the CPU, and more for the GPU. To be economically viable, Intel may have to make GPU-less CPU's, just for this. So maybe 15W for the CPU, and 35W max for the GPU part ?
https://forums.guru3d.com/data/avatars/m/254/254725.jpg
I wonder what sort of price range this will be in?
data/avatar/default/avatar23.webp
So, no more Intel integrated graphics card on solutions using a discrate card? Finally... Otherwise I do not see what is new at all.
https://forums.guru3d.com/data/avatars/m/258/258664.jpg
Cool stuff.
data/avatar/default/avatar36.webp
So they are doing what they denied 2-4 months ago and they are "gluing" it together, interesting to see the outcome.
https://forums.guru3d.com/data/avatars/m/152/152580.jpg
This should mean a significant increase in interest from developers for using GCN specific solutions and the end of the era, optimization mainly for NV.
https://forums.guru3d.com/data/avatars/m/63/63170.jpg
BigMaMaInHouse:

I think that the reason they are doing this is because Apple gonna order a lot chips from them, already they are using Intel CPU + AMD GPU so this is expected, moreover this way they are keeping Apple from going Full AMD SOC(GPU/CPU=APU).
Good point. Intel would do anything to keep Apple as a customer....
https://forums.guru3d.com/data/avatars/m/271/271903.jpg
So now I can run my Linux on Windows on an AMD-Intel chip. I must remember to take a coat when time comes to go to hell, it must be frosty down there.:):):):)
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
Keep in mind, the AMD GPU is still technically discrete, and Intel is still using their own integrated graphics in this configuration. So far, this configuration seems to be specific to just mobile platforms. This really isn't that special once you realize it's nothing more than a HBM2-powered GPU that works with Intel's EMIB platform.
https://forums.guru3d.com/data/avatars/m/267/267641.jpg
Dont be too excited, it could be other OEM not for normal mortals stuff like Intel Iris GPU.. I tried to build 35W TDP machine with decent GPU and i found out that i cant buy any of Iris GPU bellow 65W, because they are not for normal mortals, but only for club of intel OEM friends..
https://forums.guru3d.com/data/avatars/m/259/259654.jpg
schmidtbag:

Keep in mind, the AMD GPU is still technically discrete, and Intel is still using their own integrated graphics in this configuration. So far, this configuration seems to be specific to just mobile platforms. This really isn't that special once you realize it's nothing more than a HBM2-powered GPU that works with Intel's EMIB platform.
The only reason this is done is for reducing the total footprint. It's a partnership on the board level, not on the chip level.
https://forums.guru3d.com/data/avatars/m/156/156133.jpg
Moderator
kastriot:

Why not nvidia chips? they consume less power/perf, prolly because nvidia don't give a crap about intel 🙂 P.S.(Just kidding all this is on component level nothing new, same ol..)
Well for a deal, both parties have to be on board for it haha.
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
Aura89:

Is this supposed to be a picture of it? Because if so, i'm not seeing anything new here....just a CPU and a GPU, not in the same package.
Yes, pretty much. As I stated before, this isn't really anything to freak out over. Sure, it's a little interesting how Intel is working together so closely with AMD, but Intel isn't replacing their own GPU here.
https://forums.guru3d.com/data/avatars/m/56/56686.jpg
am I the only one that thinks this "dont compute" intel and amd making a product in collaboration? did hell freeze over and I miss it?
schmidtbag:

Yes, pretty much. As I stated before, this isn't really anything to freak out over. Sure, it's a little interesting how Intel is working together so closely with AMD, but Intel isn't replacing their own GPU here.
sure intels gpu made strides but they still suck even for basic usage imo, but debatable on what one calls basic usage too
data/avatar/default/avatar35.webp
People seem worried about high power comsumption for vega, but this isnt the case. Power is only high under the desktop environment where people expect performance and really cranking up clock speeds and voltage. But if you undervolt vega, it performs extremely well for perf/watt. Its also not a 4096 shader die, likely in the area of 1024.
https://forums.guru3d.com/data/avatars/m/245/245459.jpg
Aura89:

Is this supposed to be a picture of it? Because if so, i'm not seeing anything new here....just a CPU and a GPU, not in the same package. https://ae01.alicdn.com/kf/HTB11XmeRFXXXXXBXpXXq6xXFXXXQ/For-Dell-L421x-font-b-Laptop-b-font-motherboard-mainboard-with-i7-3667U-CPU-NVIDIA-font.jpg
The difference is they're putting it all in one small package, under a single IHS, to be cooled by a single heatsink. That's what this diagram is showing, showing the evolution from now (on left) to future (on right). (I'm referring to Image #1 in this article).
https://forums.guru3d.com/data/avatars/m/248/248994.jpg
Robbo9999:

The difference is they're putting it all in one small package, under a single IHS, to be cooled by a single heatsink. That's what this diagram is showing, showing the evolution from now (on left) to future (on right). (I'm referring to Image #1 in this article).
It also has that "Embedded Multi-Die Interconnect Bridge", which must be something new Intel came up with to connect dies. If the interposer AMD is using to connect the Vega and the HBM2 is one of the reasons why they can't make enough of the modules to ship to the custom video card partners, I hope this Intel solution is much better. Although if it is, I'm sure Intel covered it with 1000 patents to make it useless for everybody else.
https://forums.guru3d.com/data/avatars/m/63/63170.jpg
Robbo9999:

The difference is they're putting it all in one small package, under a single IHS, to be cooled by a single heatsink. That's what this diagram is showing, showing the evolution from now (on left) to future (on right). (I'm referring to Image #1 in this article).
This. Its a lot smaller footprint than having the two chips (and memory) on the PCB, and makes for a much simpler motherboard layout and cooling
Kaarme:

It also has that "Embedded Multi-Die Interconnect Bridge", which must be something new Intel came up with to connect dies. If the interposer AMD is using to connect the Vega and the HBM2 is one of the reasons why they can't make enough of the modules to ship to the custom video card partners, I hope this Intel solution is much better. Although if it is, I'm sure Intel covered it with 1000 patents to make it useless for everybody else.
Its still an interposer technically, its just only used to connect the two chip interfaces, rather than connect the two chips entirely. its a LOT cheaper (for the interposer cost), but its technically more difficult to get right (more expensive to implement). Intel probably owns all the patents, yes, which sort of screws it up for the rest of the world...unless they pay Intel, or have an exchange agreement with a supplier (AMD?).