6-Core Intel processors going mainstream in 2018 with Coffee Lake

No, it won't make sense. Intel already has i7 "extreme" brand.
Hopefully Intel being a nice guy and adding 2 cores for higher tier i5.

2018 should be about time we had "quality performance" 6 cores without paying $400.

 

Same! Still runs everything I throw at it reasonably well.


Really though I've been toying with the idea of building a dual CPU xeon rig.

 

OK first of all how could you start your day without coffee especially if it's a 6 cores one:coffee2:
now my reading between the lines, announcing a new socket, a 6 core going main stream, guess this is Intel own way of saying: set front shield to full, prepare for impact :bash:
is AMD pulling a rabbit out of the hat with Zen or just another over-hype :nerd:

 

Or if you wana make the naming schemes mean anything..

I3, 3 cores. I5, 5 cores. I7 7 cores. Whoop whoop

 

What everyone needs to understand is that it doesn't matter if intel drowns the cheerful AMD Zen parade with their coffee. What matters is that we may finally have a great amd cpu that offer both great IPC and Multithread that would have a reasonable price. If intel finally reacts to this and offers better prices for the mainstream, then we are all winners.

If Intel wanted to, they could very easily fit 8 cores on an i5 or i7 by removing the iGPU and then sell it for 200$ for a great profit margin... But they they're not a charity organization and will squeeze you as far as you can be squeeze. Brand loyalty is stupid and even the current underdog will become evil if things somehow switch between amd and intel.

 

does iGPU count as a core? :eyes:

 

Currently using the IGP on my i7 with my projector running it at 120hz although it's only 720p but works perfect in 3D and the main reason I do this is all the ports on my 980 are used up with my VR stuff and my monitors so yeah some people do use IGP's and let's not forget the promise of DX12 which could possibly give us access to the IGP to use along side the main GPU, what they could probably do is produce two versions.

 

knowing intel they will probably come up with an i9 with 6c/6t and i11 with 6c/12t at even higher price points and keep the lower models the same

 

That would make no sense.

The i7 series already has 6c/12t CPUs on the higher end boards going as far back as LGA-1366.

They even have an 8c/16t CPU on the LGA2011-3 platform.

 

A 10c/20t as well.

 

This is so when AMD releases there 8c 16t for 300$ Intel doesn't need to drop it's prices it just gives you 2 extra cores for the same price.

It will be an i7 6 core 12t for 400$ K versions to go against AMD's 8c 16t Black Edition versions.

Intels 6c 12t chip will beat AMD's 8c 16t but cost 100$ more.

For a gaming rig they will be evenly matched while the intel chip will beat it at single core performance.

 

TOCKs become a lot more than expected. Weren't they saying recently, that from tick-tock they are going to tick-tock-tock cycle? Now it's a tick-tock-tock-tock(that's a strange clock ).

 

So i was about to upgrade my 3770k to 6700k but then saw this article, but looking at the chart i would rather wait till cannonlake for process cycle or icelake for architecture improvement.

Well all this depends on how Zen performs.

 

Intel were rumoured to be having issues with Cannonlake, just like they did with Broadwell, in turn prompting the release of Haswell Refresh. This seems to be confirmed since it will only be on the low powered processors, where the low clocks is a benefit for power use reduction.

It will be interesting to see how the Zen 8-core, 16-thread stacks up against coffee lake 6-core, 12-thread. This very much includes the consideration of price.

 

Except in 2018 they will still cost an arm or a leg if I know Intel, unless some competition will kick in.
The performance progress of intel cpus has been so slow it is getting almost to an halt.
I used to upgrade regulary, but now I still have my core i7 Nehalem from 2009 at 3GHz.
I remember when the first engineering samples and unofficial benchmarks of Haswell leaked out, I saw some comments on a discussion forum about not getting our hopes high on Haswell.
Then an Intel engineer answers very angry stating that Haswell would be indeed faster.
Oh well Haswell came out and then some other cpus after that and here we are standing still.
Unfortunately my hobby demands a CPU with high IPC and frequency and high core count (for chess and checkers engines, and generation of endgame databases that take many months to complete).
I'm not interested in SIMD and FPU and graphics performance so I'm an user that was hit the most by this lack of cpu progress.
If someone says it's because of technological limitations that Intel can't go forward!
Well that is not an explanation, very little is impossible if we put our minds working at it, after all wasn't Intel that in 2000 said it would have a 10GHz cpu by 2010?
I remember in the past US tech companies made almost miracles for the military.
So complaining it is technological limitations is no explanation.
It is like saying the electric car hasn't progressed beecause of technological limitations, well we know the limitations come from other camps.

 

Technological limitations is a serious problem since we are using electrons that are traveling near the speed of light through tiny silicon routes and they tend to fly off the controlled surface as we shrink those surfaces down to smaller and smaller nm's.

16nm, 14nm, 12nm, 10nm already has some leakage of those electrons so that's a huge problem that seems to have no solution.

This is why Intel has struggled to improve performance because like i said, smaller = faster but smaller & faster = more loss of control of those electrons to a point where the gains are offset by the losses.

 

Funny how Intel forgot themselves that explanation in 2000, when they promised in 2010 they would have a 10GHz cpu.
Anyway who said we must continue the silicon/electron path?
And about eletrons traveling near the speed of light, really? We must live in different universes with different universal constants.
In the universe I live in electrons can travel at incredible slow speeds (in the order of some mm or cm per second), wich shouldn't be confused with signal speed wich by the way is also far from light speed.
Anyway and again who said we must persue that silicon/copper/electron path?

 

The issue with that is that CPUs increased the number of physical cores on a di which effects CPU clock speed. Look at the top of the line Haswell and Broadwell E chips and their clock speed vs a Quad core Haswell and Broadwell chip. Maybe if we stuck with single and dual core cpus then we may have gotten to 10ghz. But I think Intel was just predicting a 10ghz CPU, not promising one.

 

If you allow me making a metaphor, let me say it this way: it's like planes in the era of piston engines, we need to let go the piston engine use in planes and start using the jet engine. It is clear the piston engine won't get us much far ahead. How about some inovation? I mean with so much money power if there is anyone that could do it, it should be Intel no?. How about that Terahertz transistor they said they had quite a few years ago?
But not only that, the one thing that gets me the most is the way Intel milks their tech so much, in every "generation" of supposedly much better products keeps charging unbelievable high prices, for almost nothing added from to the previous "generation". That's what we get by lack of competition.
Not to mention the devious ways of Intel in the case of AMD for bribing and threatening PC makers as well as retailers to not use AMD cpus. Or is this supposed to be normal business, trying to kill the competition? This company leaves a lot to be desired in every way.