• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Intel to launch 6 core Coffee Lake-S CPUs & Z370 chipset 5 October 2017

The performance of threadripper and price (forgot about that) will affect the effective lifecycle of X299, which has a knock on affect on there product stack due to reason i mentioned previously.

Your thinking about it from the wrong perspective. Intel wants as much money as possible. It is Intel's best interest to have people like him (who are sitting on the fence) to become impatient and buy into the X299 platform. Heck who knows a maybe later on they will still end up buying a coffee lake system as well. Win Win as far as Intel is concerned. For this reason, they will not want to release to much information as it could potentially cause harm to the sales of X299.

That's a lot of speculation and only one possibility. Don't forget that this HEDT, too, is meant for prosumers (where the mesh migration is the only place it makes sense) and of the line the 7800K is the absolute worst on all fronts - enter Coffee Lake 6C/12T, way better in many scenarios with its traditional ring and cache design.

I don't think Intel wants to get everyone on HEDT, and nor is everyone considering it - the minority is, is what I'm grasping from multiple large fora I visit.
 
That's a lot of speculation and only one possibility. Don't forget that this HEDT, too, is meant for prosumers (where the mesh migration is the only place it makes sense) and of the line the 7800K is the absolute worst on all fronts - enter Coffee Lake 6C/12T, way better in many scenarios with its traditional ring and cache design.

I don't think Intel wants to get everyone on HEDT, and nor is everyone considering it - the minority is, is what I'm grasping from multiple large fora I visit.
By mesh migration are you talking about mesh architect they introduced? Not sure about the benefits in applications but one of the "big" reasons it was introduced was because it scales better with more cores than the previous ring bus design.

Intel entire business model is getting you to go up a product level. That why they segment the market and place arbitrary, limitations on there product stack. It is definitely in Intels interest that he buys X299. Not everyone posts their concerns on forums but it is an actual consideration.
 
By mesh migration are you talking about mesh architect they introduced? Not sure about the benefits in applications but one of the "big" reasons it was introduced was because it scales better with more cores than the previous ring bus design.

Intel entire business model is getting you to go up a product level. That why they segment the market and place arbitrary, limitations on there product stack. It is definitely in Intels interest that he buys X299. Not everyone posts their concerns on forums but it is an actual consideration.

Yes that mesh is what I referred to. It makes perfect sense for the truly high core count SKUs.

It's a good thing he's posted his dilemma. All I can say it look at the sort of user you are and pick the fitting platform. Personally I am glad I took two steps back and looked at it again objectively, and it lead me to CFL-S while saving significant money whilst getting the exact same performance and all that I need.
 
Ah the good old days where i3 = 2c/4t, i5 = 4c/4t, i7 = 4c/8t, and i7-X = 6c/12t. So simple.
What's good in something which has kept game developers uninspired to optimize multithreading?
(because they can't make games just for PCs of enthusiasts)

Sure word processing/web surfing could do well with 2c/4t but in other uses would be time to up that absolute minimum to 4 cores with standard level being 4c/8t.
All that background stuff running in PCs sure won't be getting any lighter or less numberous.
 
All I want is 6c/12t 4ghz+ chip that slots into the Z270 platform. There should be no reason why that can't be done.

Perfect for gaming and good enough for video/photo editing.

Get that chip out at under £380 and I'll buy it in a heart beat.
 
All I want is 6c/12t 4ghz+ chip that slots into the Z270 platform. There should be no reason why that can't be done.

Perfect for gaming and good enough for video/photo editing.

Get that chip out at under £380 and I'll buy it in a heart beat.

Socket probably doesn't have the necessary number of pins for a 6 core chip.

I also doubt they designed the chip to be backwards compatible, why would you with no competition?

Coffeelake was supposed to be Cannonlake on 14nm from memory.

I hope someone dies a 14nm chip comparison in terms of IPC and performance per watt.

Broadwell (Haswell shrink) -> Skylake (Architecture) -> Kabylake (Optimisation) -> Coffeelake (Optimisation 2) -> Cannonlake (Skylake shrink)

4 years of processors (if Icelake comes out in 2019).
 
Last edited:
What's good in something which has kept game developers uninspired to optimize multithreading?
(because they can't make games just for PCs of enthusiasts)

Sure word processing/web surfing could do well with 2c/4t but in other uses would be time to up that absolute minimum to 4 cores with standard level being 4c/8t.
All that background stuff running in PCs sure won't be getting any lighter or less numberous.
Yes but it would've been better if they changed all of them at once, rather than quietly making Pentiums 2c/4t without changing Core i3, and then soon bringing in 6c/6t Core i5s and 6c/12t Core i7s whilst still keeping existing 4c/4t and 4c/8t CPUs with the same monikers. It's slower progress than we'd have had without competition and it's haphazard, which makes it unnecessarily confusing.
 
All I want is 6c/12t 4ghz+ chip that slots into the Z270 platform. There should be no reason why that can't be done.

Perfect for gaming and good enough for video/photo editing.

Get that chip out at under £380 and I'll buy it in a heart beat.

Hey rad! How's you?

I'm afraid it won't be backwards compatible, I wish! It'll require LGA1151 v2, same number of pins but likely a different pin layout.

CFL-S on GeekBench:

https://redirect.viglink.com/?forma...ttps://browser.primatelabs.com/v4/cpu/3208482
 
Last edited:
How much do intel get from each socket change, per board sold?
I can't see it being that much.
AMD and Intel sell the chipsets to motherboard vendors. Someone was saying they can cost upto $40. Intel also has the intel NIC that some boards come with. Potentially a few MB patents that they might hold on to.

Its a fair chunk of change.
 
AMD and Intel sell the chipsets to motherboard vendors. Someone was saying they can cost upto $40. Intel also has the intel NIC that some boards come with. Potentially a few MB patents that they might hold on to.

Its a fair chunk of change.

In that case I can see why they keep changing the bloody socket!
 
It also keeps motherboard manufacturers happy as Intel is basically creating demand for them.
I'm not so sure tbh.
It depends how much work they need to do to incorporate the new chipsets. If its a lot of work then I don't think 12 months is enough to maximize the product life cycle of a motherboard (Its enough to make money but not the maximum for the given investment). If its as simple as popping the old chipset out and sticking the new one then use they would be quite happy.
 
I'm not so sure tbh.
It depends how much work they need to do to incorporate the new chipsets. If its a lot of work then I don't think 12 months is enough to maximize the product life cycle of a motherboard (Its enough to make money but not the maximum for the given investment). If its as simple as popping the old chipset out and sticking the new one then use they would be quite happy.

Potential difficulties are immaterial; imagine a board maker not making any boards for 2-3 generations... It's good money for both board makers and Intel, I imagine.
 
Potential difficulties are immaterial; imagine a board maker not making any NEW boards for 2-3 generations... It's good money for both board makers and Intel, I imagine.

Just corrected that for you :D

If you go back to NF2 days, we had the same socket and chipset for some time. In fact we also had DFI as a board manufacturer then. They went not long after x58, maybe it was because of the short life of mobo's and the investment needed ?
 
Potential difficulties are immaterial; imagine a board maker not making any boards for 2-3 generations... It's good money for both board makers and Intel, I imagine.
Your comment just convinced me that this is merely Intel trying to make as much money from enthusiast as possible, and has little to no benefits to MB vendors.
Most people don't upgrade multiple times in a generation, yes you have the enthusiast that will buy every new generation of processors, but these people are a tiny minority; and the profit from those people would not be significant enough to offset the cost of designing a new MB, as well as "scraping" (using that term loosely) a previous board design.
Most people won't be upgrading their machines for at least 2 years (Would probably upgrade between 3-5 years). With such a time frame the average person will need to buy a new motherboard because there's would be obsolete even with AMD time frame.
 
Back
Top Bottom