- Mar 10, 2006
- 11,715
- 2,012
- 126
It probably won't be any better though. Also seems like it will still only have 8 MB L3 too.
On what do you base that cache claim? I don't think there's evidence either way.
It probably won't be any better though. Also seems like it will still only have 8 MB L3 too.
On what do you base that cache claim? I don't think there's evidence either way.
Die size. Isn't it 140ish vs 120 or so for the Kaby Quad? The dual is like 90 something. Plus it makes sense with the L3/core cut on Skylake Server which is pretty much confirmed.
However one can hope.There's no official leak about CFL being hexa-core, except a site named PCWatch's own speculation due to what they 'heard'.
There's no official leak about CFL being hexa-core, except a site named PCWatch's own speculation due to what they 'heard'.
There's no official leak about CFL being hexa-core, except a site named PCWatch's own speculation due to what they 'heard'.
i got an interesting whatsapp message today:
CFL-62 design "tape-in" last week (a term i'm not familiar with, i know of "tape-out", but not "tape-in"), first desktop cpu with 6 cores, uses KBL architecture. product due 2018.
6700K will last until the end of this console generation.
Assuming PS5/XB2 are slated for 2019-2020, I would want Ice Lake/X or it's refresh. Coffee Lake's architecture will be 2.5-3 years old by the time it launches. It will fall into no man's land territory on the desktop side with new Ice Lake architecture not far away at that point.
I actually have a feeling that based on how much slower Intel will update new architectures, 6700K users could even skip Ice Lake and wait for "Ice Lake 2" 2021's architecture. I went from Sandy to Skylake which is a 2 generation architectural jump (Sandy -> (1) Haswell -> (2) Skylake. The "same" jump should now take ~ 6 years since Intel moved away from Tick-Tock.
Throwing in more cores for marketing reasons (and maybe someday actual performance reasons) is also clearly on the agenda.
Honestly the real reason they are doing Coffee Lake is because they want to have something new but didn't want to spend the $$$ to get the 10 nm fab capacity needed. Plus to burn some more 14 nm wafers.
I actually have a feeling that based on how much slower Intel will update new architectures, 6700K users could even skip Ice Lake and wait for "Ice Lake 2" 2021's architecture. I went from Sandy to Skylake which is a 2 generation architectural jump (Sandy -> (1) Haswell -> (2) Skylake. The "same" jump should now take ~ 6 years since Intel moved away from Tick-Tock.
The big if is what they mean by white dots on the process roadmap for 10nm. Will the staggered introductions be such that we'll see 10nm cores introduced well into the 7 nm cycle? Like we might see 2nd "half gen" of 7nm alongside 5th "half gen" of 10nm?
I just realised the ideal target for 6 core Coffee Lake- iMacs. Coffee Lake plus a Vega GPU will probably be faster than the current Mac Pro.
Yup.
It will gain bonus points if it's drop-in compatible in Z170 boards.
For products that don't need to be moved to the shrunken nodes, Intel's strategy seems to be to keep them on the larger geometry nodes (with improved transistor performance) until the next shrunken node becomes economical.
Mobile is all about efficiency. A 6 core will have the same perf as a quad or even dual when everything is limited by TDP.HEDT has a price floor. It's a big expensive platform with quad channel memory, tons of PCIe lanes, and support for multi-socket systems (even if this is disabled in consumer parts). All of this adds transistors on the CPU, and complexity to the motherboard. They can't get the price as low as if they made a dedicated 6 core consumer chip.
Anyway, the big win will be in laptops. 6 core Macbook Pro, anyone?
Mobile is all about efficiency. A 6 core will have the same perf as a quad or even dual when everything is limited by TDP.
(the improvements made in 14nm+ for example are not trivial and certainly not the types of improvements you'd get just from process maturity)
Please post your source for this information. All we've seen so far is that the 7700k's baseclock is 200Mhz higher than the 6700k. The turbo is 300Mhz higher. This is not excting, since every Skylake CPU ever made can clock to these speeds.
I've seen some rumours that the 7700k can clock to 5.1Ghz on air etc, though we all saw the same nonsense rumours about the 6700k before release also.
So, please post your source showing the 'non trivial' gains from 14nm+.
What started a simple pet project to get comparable Kaby Lake and Skylake systems in house quickly became a riveting dive into the performance advantages of Intel’s newest architecture. I did not expect the gap between these two most recently platforms to be as wide as it was or as ubiquitous through my testing.
For example, looking at something like H.264 video encoding, I was fully expecting a ~5% performance advantage that was in line with the base clock improvements from the specifications we showed you on an earlier page. Instead, we saw well over double digit gains in some areas, proving that clock speed, thermal improvements and technologies like the updated Speed Shift were combining into a “better than the sum of its parts” result. Other noted surprises were the gaming benefits (over 30% faster in Overwatch) and an 18% edge on POV-Ray, a heavily multi-threaded rendering engine.
With 14nm+, Intel changed the fin height (made the fins taller, which improves drive current) and made some other materials changes ("improved transistor channel strain"):
Intel made no such changes during the lifetime of 22nm.
Anyway, Kaby Lake is already out for notebooks and reviewers showing big performance gains at a given power consumption thanks to the new transistors (since top frequency is higher and the more efficient transistors mean that the chips can stay at higher frequencies for longer).
https://www.pcper.com/reviews/Processors/Intel-Kaby-Lake-Performance-Surprising-Jump-over-Skylake
Whether this translates into significantly higher OC capability on desktop remains to be seen, but +300MHz in single core turbo and probably +400MHz in all-core turbo at the same power consumption should verify that 14nm+ offers meaningfully better performance than 14nm.
Intel marketing slides?.....Speedshift? Clockspeeds obtainable on every Skylake? I'm talking from a desktop point of view here.
Again, pretty much every unlocked skylake out there can clock to 4.6, 4.7Ghz easily, so Kaby being able to do 4.5Ghz turbo (on one core?) is pathetic.