Laptops. Intel wants to absorb the dollar content that the dGPU makers have with more powerful iGPUs.
Intel seems deadset on increasing its iGPU performance. I have no idea why, but it's quite interesting. I really am interested as to where it takes us.
That chart has no majority. The plurality goes to 2 CPUs, which is almost certainly dominated by laptop CPUs rather than i3 desktop CPUs and below.
They need to compete against those 6-core Apple SoCs, right? :awe:
[/sarcasm]
Laptops. Intel wants to absorb the dollar content that the dGPU makers have with more powerful iGPUs.
There are limitations, but before they get to the technical limitations they will run into the artificial limitations. Artificial meaning it is created by choices of the engineers, such as being pad limited for power distribution, or periphery limited for IMC bandwidth, etc.Out of curiosity, aren't there limitations on how small they can make the die. Can they make a 50mm cpu for example?
Let's see if AMD can ruin their party with Zen based APU &/or Excavator plus HBM2 btw will Intel continue to go eDRAM/L4 cache route or will they ever incorporate HBM/HMC anytime in the near future? The last part an open question to anyone who's got some deep(er) info about their product mix.Laptops. Intel wants to absorb the dollar content that the dGPU makers have with more powerful iGPUs.
+1, its another place where they can sell silicon that takes advantage of their process tech in terms of performance/W too.
Let's se e if AMD can ruin their party with Zen based APU &/or Excavator plus HBM2 btw will Intel continue to go eDRAM/L4 cache route or will they ever incorporate HBM/HMC anytime in the near future?
Nope not first but I do expect Zen APU's to compete with Cannonlake, like 6~12 months after the latter's launch depending on how GF does their job. As for rumors of Excavator at 14nm, it was slated to be H2 of CY 2016 IIRC.Do you have any evidence that the first Zen APUs will have HBM/HMC? These are expensive technologies and I don't think you will see them in client CPUs anytime soon.
Nope not first but I do expect Zen APU's to compete with Cannonlake, like 6~12 months after the latter's launch depending on how GF does their job. As for rumors of Excavator at 14nm, it was slated to be H2 of CY 2016 IIRC.
You mean desktop Excavator?There will not be a 14nm XV; the 2H 2016 XV parts will be on 28nm.
You mean desktop Excavator?
That's the problem. They're assuming Intel means 4770k+ when they say "Extreme gamer".
We have NO clue what intel means. This is just enthusiasts thinking that everyone is talking specifically to them/about them/etc. and thinking that because Intel says Extreme gamer, it means the bracket THEY BELIEVE is extreme gamer. Especially when they can further shape that into saying that Intel should now provide a hexacore as the new i7 chip, because they want that. Not because it's anything based in reality.
Like I've said before, enthusiasts are completely out of touch with the rest of the market. I've only just returned to enthusiast level purchases, but still I understand that no product is made with me first in mind from intel. They're thinking about mobile first, as they should. I LOVE the strides they make in lowering power consumption and making faster chips fit in smaller and smaller devices.
Most of the tasks people are talking about, if you really care about those tasks, you'd get the HEDT platform and not worry about the tiny incremental improvement in the latest platform. But you don't really need those tasks if you're complaining and don't have the HEDT platform. You just want it cheaper/newer. And of course we all do. But it's not going to happen, and it's perfectly reasonable as to why.
Edit: This whole situation reminds me of the Nintendo employee that got fired while explaining to neogafers that their small userbase wasn't enough to justify Nintendo localizing games for them.
Exactly. Essentially its about the feeling of entitlement rather than any base in reality.
I agree with Ashraf (the author of the article we are discussing) in that the argument exists to expand core count at 10nm because of how tiny a quad-core die will be at that point. And doubling CPU cores does not double the silicon die area, probably will only be a 15% adder to go from 4 core to 8 core at 10nm.
1. 10nm will be more expensive than 14nm
2. iGPU die size will increase
3. Intel will integrate the north bridge
4. they may need to integrate the eDRAM (TSV)
With all those above, the die size will be close to current Skylake but at more expensive 10nm process. Adding 4 more cores will make it even more expensive (bigger die, lower yields etc).
Dont believe we will see 8-cores for the client at 10nm.
1. 10nm will be more expensive than 14nm
2. iGPU die size will increase
3. Intel will integrate the north bridge
4. they may need to integrate the eDRAM (TSV)
With all those above, the die size will be close to current Skylake but at more expensive 10nm process. Adding 4 more cores will make it even more expensive (bigger die, lower yields etc).
Dont believe we will see 8-cores for the client at 10nm.
Intel will integrate the north bridge
It's getting proportionality cheaper for intel to throw in 2-4 more cores each generation, even with iGPU increase.
And yet you are just like everyone else, I listed an array of reasons more cores are better, yet that list is ignored because there is no good counter argument to it.
The only argument I ever seem to get it 'well you should just pay for HEDT because you happen to use your CPU too much'.
DX12 btw will more likely lower CPU usage than increase it. And for encoding there is quicksync.
So what do you think the engineer in the OP is working on?
By north bridge you mean PCH? Aka former south bridge.
I have said before, I believe its for server parts.