I’m pretty sure either somebody called this out as likely in one of the episodes of the news, or the complete opposite where this was speculated in the media and the guys shut it down as near-impossible given their history.
What to make of this? Nvidia has to be a little worried. I’d be much more excited to see Intel and AMD team up to tackle the AI/Deep-learning applications as well.
Definitely interesting. I don’t believe that this is immediate cause for worry for Nvidia, but this really could hurt them in the long run. Right now this is really a new product as thin and lights really haven’t had powerful graphics ever.
Anyway, I wonder if this will be Polaris or Vega based, because AMD has their new Raven Ridge with Zen and Vega.
In May there was a rumor that Intel bought the license for some of AMD’s graphic chip design. AMD shot down the rumor. I wonder if this is what the rumor was based off of.
Most likely. Intel really didn’t buy any licenses, they just ordered a custom Vega graphics silicon from AMD…
Curious how this will do…
Last time this rumor spawned, it turned out sombody working at Intel has the same last name as AMDs new GPU architecture (Vega that is).
Now for the “new rumor”:
I find this unlikely. AMD has a very competetive CPUs and is the only other supplier of graphics hardware for notebooks. This deal would not benefit them in any way as they would be giving their stronger graphics away to become an argument to buy the Intel notebooks.
official intel newsroom article:
not a rumor. they will be packaging amd discrete graphics with intel cpus, not to be confused with intel cpus having an igpu embedded. two seperate devices on the same board.
It’s official…
Why tho?
What can they win?
I seriously hope there is something I am missing here… else there are no benefits in sight.
"The MCM should enable devices as thin as 11 mm.“
I have always dreamed about 11mm lapt… no I haven’t…
” enables a new breed of ultra portable notebooks"
it’s just a mobile thing… small mobile thing.
so it is meh?
I see four possible non-direct tech related reasons:
that, yeah…
Nvidia have become too large even for Intel’s taste…
The stock market liked it:
Strong feeling these are going to end up in New Apple products…
Also mildly related and funny
Wasn’t there some buzz about Intel developing chips with a socket part for the graphics processing to be added later? With rumours that they were doing so to adopt some sort of AMD chip to handle the graphics? I’m almost certain this came up in the L1T news at some point.
What i don’t understand is why a Giant like Intel with $60 Billions revenue has to hire a smaller company to make chips for them. The only guess I have is they don’t want to hire skilled engineers because they would have to share some of their wealth.
because almost every part of the gpu has a patent around it making it almost impossible to make a competing product
AMD aren’t directly competing with themselves by doing this. The Raven Ridge APUs that they announced the recently are going up against Intel’s ULV chips.
These chips are going to be Core-H series (around 45W). AMD don’t have any APUs in that area right now, so it makes sense for Radeon to be working with Intel on this. Means you don’t need a dedicated Nvidia GPU with these chips.
intel is falling behind in the gpu market, and nvidia probably makes it a pain in the ass to integrate their discrete gpus. if their linux presence has said anything, it’s said nvidia would rather make someone go around their ass to get to their elbow in terms of efficiency and getting something working. Just look at the nvidia linux drivers. Sure, they work. But they work as a hacky blob that is not integrated into the kernel. They also dont provide any documentation so that nouveau (the open source nvidia linux driver) can progress. They also began signing their firmware after the 7x series so nouveau can’t do anything with nvidia firmware blobs either.
Furthermore:
Goodbye broken-as-shit nvidia mobile linux drivers. With intel and AMD combined, open source mesa can finally add proper GPU switching rather than nvidia forcing linux users to use only the nvidia gpu with an x passthrough with no swapping or power saving. This also means not having to rely on eglstreams and better wayland support on the intel/amd linux front.
This also probably benefits apple as well indirectly, seeing as mesa is open source and can provide a method of getting things to work. Not saying apple would take any cues from open source, but we all saw how metal magically appeared shortly after apple dropped off the vulkan groups list. Also, both intel and amd are easier to work with than nvidia.
lol because they suck in the GPU department?
To be serious though… patents really, there’s not really any way to design a GPU compatible to todays software without breaking a shitton of patents. It’s the same issue as with x86 CPUs.