Intel's 8th Generation Core Family - Coffee Lake (LGA 1151, 6C/12T)

Where do you expect Core i7-8700K's Turbo to land?

  • 3.8/3.9 GHz

    Votes: 0 0.0%
  • 4.0/4.1 GHz

    Votes: 3 23.1%
  • 4.2/4.3 GHz

    Votes: 6 46.2%
  • 4.4/4.5 GHz

    Votes: 3 23.1%
  • 4.6/4.7 GHz

    Votes: 1 7.7%

  • Total voters
    13
  • Poll closed .
Well that's a bummer. I was looking forward to getting some CFL details this morning (as was most of the tech world apparently). Guess we're all stuck waiting a bit longer. Granted, maybe this means we'll see Z390 availability with CFL launch now.

Why? The z370 rumors are thick, fast and consistent. If anything it's even worse now, the z370, instead of being on shelves Aug 2017 could be Sept or even Oct 2017.
With a replacement z390 potentially less than 5 months later........
 
Why? The z370 rumors are thick, fast and consistent. If anything it's even worse now, the z370, instead of being on shelves Aug 2017 could be Sept or even Oct 2017.
With a replacement z390 potentially less than 5 months later........


I think z390 will be for cannonlake launch, and cannonlake desktop will probly be in a year because they will start will low TDP. mostly like Broadwell... actually no desktop for broadwell because the shrink was hard and I except cannonlake to be hard but more feasible.
 
God it upsets me how expensive really high end ddr4 is. I could get 32GB (2x16, for itx) quality gear so cheap comparatively, About a year ago..

When is this crap gonna end?

In the next recession. Until then, brace yourself for higher prices.
 
As others have stated, it's possible that the chipset isn't really changing beyond getting rebranded. That said, VRM specifications can and do sometimes change. We've seen this before. Several Intel motherboards had to be of a specific revision to support newer CPU's. NVIDIA's 780i SLI chipset was virtually identical to the 680i SLI chipset it replaced aside from the nF200 chipset being added to it for PCIe 2.0 compatibility. The VRM specifications had to change as well because Intel altered the CPU specifications so that newer CPUs wouldn't work on the older 680i SLI boards. Don't forget Socket 370. There were older socket 370 motherboards that only supported Celeron CPUs, then later socket 370 motherboards that supported Coppermine Pentium III CPUs. Then there is LGA 2011. We've seen multiple versions of this socket, yet they are electrically different.

Allot of people accuse Intel of being "the bad guys" for changing sockets too often. The reality is, we've had issues that stem from not changing the sockets often enough on the AMD side. Microcode, BIOS issues, and TDP increases or voltage specifications changes prevent some motherboards from using newer CPUs anyway.
Don't I know this all too well, particularly on Socket 370...

Take the Abit BP6, for instance. When I picked mine up so many years ago: "Oh, hey, maybe I can drop dual Tualatins into this!" NOPE, SO MUCH NOPE.

The board needs extensive power delivery modification (read: unsoldering old capacitor and VRM components, replacing them with new ones) AND a pair of PPGA to FC-PGA socket adapters just to run Coppermine P3s, and hardly anyone's managed to run a single Tualatin (which changed the Socket 370 spec again to FC-PGA2) on an extensively hotrodded BP6, with NO confirmed reports of dual Tualatins in SMP. If anything, the fact that enthusiasts went to such lengths to try and do that on a BP6, on a BP6-centric forum at that, says something about its following.

There's also another issue we have with modern motherboards: the current PCHs do much less than the northbridge/southbridge chipsets of old because what used to be the northbridge functionality is now part of the CPU, things like the memory controller and PCIe lanes.

Thus, a lot of the motherboard's functionality is largely dependent on what CPU you insert to begin with, and this holds especially true on the mess that is the X299 platform, with Kaby Lake-X hamstringing the whole thing to "mainstream" desktop specs (only dual-channel RAM and 16 PCIe lanes) and lower-end Skylake-X parts only having 28 PCIe lanes instead of the full 44 lanes.
 
I think z390 will be for cannonlake launch, and cannonlake desktop will probly be in a year because they will start will low TDP. mostly like Broadwell... actually no desktop for broadwell because the shrink was hard and I except cannonlake to be hard but more feasible.

There is no Cannon Lake desktop. Coffee Lake desktop and then straight to Icelake desktop.
 
Lackluster presentation, for me at least, was expecting some concrete launch dates for the desktops CPUs in the worst case scenario. On the bright side this just gave me the nudge to go on x299. Gotten the mobo yesterday, just need to make up my mind on the CPU next... Decisions decisions...
 
Im curious as to how will the new 15w quads stack up against the older skylake/kaby HQ CPUs?

The Skylake/Kaby HQs seem to have lower max turbo speeds but higher TDP (45 vs 15w)
 
The 8th Gen i7 packaging shows that Coffee Lake will only do Turbo 2.0, not Turbo 3.0
 
Don't I know this all too well, particularly on Socket 370...

Take the Abit BP6, for instance. When I picked mine up so many years ago: "Oh, hey, maybe I can drop dual Tualatins into this!" NOPE, SO MUCH NOPE.

The board needs extensive power delivery modification (read: unsoldering old capacitor and VRM components, replacing them with new ones) AND a pair of PPGA to FC-PGA socket adapters just to run Coppermine P3s, and hardly anyone's managed to run a single Tualatin (which changed the Socket 370 spec again to FC-PGA2) on an extensively hotrodded BP6, with NO confirmed reports of dual Tualatins in SMP. If anything, the fact that enthusiasts went to such lengths to try and do that on a BP6, on a BP6-centric forum at that, says something about its following.

There's also another issue we have with modern motherboards: the current PCHs do much less than the northbridge/southbridge chipsets of old because what used to be the northbridge functionality is now part of the CPU, things like the memory controller and PCIe lanes.

Thus, a lot of the motherboard's functionality is largely dependent on what CPU you insert to begin with, and this holds especially true on the mess that is the X299 platform, with Kaby Lake-X hamstringing the whole thing to "mainstream" desktop specs (only dual-channel RAM and 16 PCIe lanes) and lower-end Skylake-X parts only having 28 PCIe lanes instead of the full 44 lanes.

I couldn't agree more. I hate the Core i7 7740X. It's stupid. That said, mine does 5.0GHz plus on the test bench and that's pretty awesome.

And it shouldn't surprise because this is just a skylake/kabylake chip with 2 more cores (and a new fucking motherboard...fuck you Intel).

In fairness, we don't really know if a new motherboard was necessary or not. I suspect the VRD specifications require an update and the difference is less about the chipset, and more about how the VRM's are built for Skylake and Kaby Lake quad core CPUs. It's not at all different from Athlons with a 130watt TDP being unable to run in a motherboard that was only spec'ed for 95w CPUs.
 
In fairness, we don't really know if a new motherboard was necessary or not. I suspect the VRD specifications require an update and the difference is less about the chipset, and more about how the VRM's are built for Skylake and Kaby Lake quad core CPUs. It's not at all different from Athlons with a 130watt TDP being unable to run in a motherboard that was only spec'ed for 95w CPUs.
So update the motherboard requirements, not the socket. Or, heaven forbid, design the CPU around the previous standard and/or future proof your standards just a tiny bit.
 
I'm confused, isn't that exactly what they did?
The socket may be the same but the mobos aren't equipped to run the CPUs.
I thought the socket has changed, but if they did to the extent that none on the boards on the market supports it, same difference.
 
Since I couldn't wait anymore - won't be waiting till "ffall" without a proper PC, and don't want to invest into old platform, I just went with Ryzen 1700 with that new EKWB aluminium loop and Gigabyte Gaming 5. Should be building new machine during weekend :)
 
Small and cash starved AMD must be a magic company then, relative to Intel.

AMD went quite quickly through FM1, FM2, FM2+ and now AM4. And dont expect AM4 to last longer than the first 14nm refresh.
 
AMD went quite quickly through FM1, FM2, FM2+ and now AM4. And dont expect AM4 to last longer than the first 14nm refresh.
Those are different market segments. They moved much more slowly for the mainstream desktop CPUs. Even then they've often kept forward compatibility.
 
Lackluster presentation, for me at least, was expecting some concrete launch dates for the desktops CPUs in the worst case scenario. On the bright side this just gave me the nudge to go on x299. Gotten the mobo yesterday, just need to make up my mind on the CPU next... Decisions decisions...

Same here. Just pulled the trigger on a 4.9 7820x from SL. :) was eyeing a 4.9 7800x but they sold out after yesterday's presentation (go figure).
 
AMD went quite quickly through FM1, FM2, FM2+ and now AM4. And dont expect AM4 to last longer than the first 14nm refresh.

Yeah and almost no one on here ran a FM socket of any kind. AM4 is good for 7nm there is no 14nm refresh on AMD Zen+ is either a 2018 Q4 or 2019 Q1 product. Deflect all you want but people on the Intel side are getting hosed by this needless chipset change.
 
Everyone bragged about how well Bloomfield, Sandy, Ivy, etc held up for 5+ years but now suddenly everyone wants to upgrade their Ryzen processor within 2 years. Maybe for workstation folks who want to move from 8C Ryzen chips to maybe 10C or 12C down the road? But for everyone else I just can't see it. I assumed Ryzen would get attention for its value proposition but the socket debate seems to be taking the crown. It's not even something I ever thought about before. My last CPU upgrade was after 7 years, this one will make 4 years. Both of which are longer than AM4's lifespan. I've never replaced a CPU in less than 4 years.

If you're the kind of person who needs to upgrade their CPU that often, or can afford to, then it seems like Intel would ultimately be a better choice since they are higher performing chips... right? Otherwise why else would you need that kind of up-to-date performance & core counts. Bizarre.

I suppose it's the easiest target in the AMD vs Intel debate, which is why people are hammering it so hard.
 
Everyone bragged about how well Bloomfield, Sandy, Ivy, etc held up for 5+ years but now suddenly everyone wants to upgrade their Ryzen processor within 2 years. Maybe for workstation folks who want to move from 8C Ryzen chips to maybe 10C or 12C down the road? But for everyone else I just can't see it. I assumed Ryzen would get attention for its value proposition but the socket debate seems to be taking the crown. It's not even something I ever thought about before. My last CPU upgrade was after 7 years, this one will make 4 years. Both of which are longer than AM4's lifespan. I've never replaced a CPU in less than 4 years.

If you're the kind of person who needs to upgrade their CPU that often, or can afford to, then it seems like Intel would ultimately be a better choice since they are higher performing chips... right? Otherwise why else would you need that kind of up-to-date performance & core counts. Bizarre.

I suppose it's the easiest target in the AMD vs Intel debate, which is why people are hammering it so hard.

Yepo, I bought 4790K, when it was launched iun 2014, and didn't bother with all the socket changes, because it was good enough. If not for my Z97 board, which died, I'd never consider the update till Ice Lake, or maybe even later. Right now, when I build my Ryzen rig, I won't be changing CPU next year for meaningless gains either (unless there is some breakthrough with gen. 10 / Zen 2) that will warrant the upgrade. But I'm going to enjoy my 8 cores for some time and only upgrade GPU, because so far, all the yearly changes in Intel lineup were only barely better than my i7.

And propably, if I'll do another upgrade in 3 years, there will be enough changes in mobo technology, that will warrrant new mobo, despite being either in Intel or AMD camp.
 
Small and cash starved AMD must be a magic company then, relative to Intel.

Well AMD has been doing it.
If Intel isn't doing it, it's either malice or incompetence.

There is no doubt in my mind that Intel probably could keep its CPUs on the same socket for longer if it wished to do so, but there are some points you guys are ignoring. AMD's been constrained by aging sockets and put out worse products because of it. Bulldozer using AM3 wouldn't allow for full support of all the C-states that the CPU supported when used on older platforms. Then you end up with motherboards that can't handle the upper echelon CPUs because their VRM's can't handle the higher power requirements of those CPUs. You had 95w only capable motherboards that wouldn't support CPUs with a 130watt TDP. AMD wasn't able to switch to newer memory technologies as quickly, or as easily as Intel did. Again, constraining a CPU to an older platform may provide for a potentially long upgrade path, but not always a pleasant one. Socket longevity created BIOS and microcode issues with various motherboards. Intel does away with all of that crap by keeping its CPU support on a given socket and chipset limited. It also limits what can be done with the newer CPU beause of those constraints.

When you are trying to improve your CPUs on a performance per watt basis, part of doing that successfully is reducing the power requirements of the platform, or optimizing that platform to the CPU. AMD doesn't do that and ends up constrained by the electrical requirements of platforms that are dated to put it nicely.

I'm not saying that Intel couldn't put Coffee Lake on existing LGA 1151 based Z270 motherboards. Although, they have done it with refresh CPUs in the past which leads me to believe there is some sort of technical reason behind it. That said, the 300 series chipsets aren't the same as Z270 as others have claimed here. It looks like it will have an integrated audio DSP and support for a new C state, (C8). It's not a huge change but Z3xx doesn't appear to be the same exact chipset as Z270.
 
That said, the 300 series chipsets aren't the same as Z270 as others have claimed here. It looks like it will have an integrated audio DSP and support for a new C state, (C8). It's not a huge change but Z3xx doesn't appear to be the same exact chipset as Z270.

Well, that makes it sting a little bit less, at least. XD
 
Well, that makes it sting a little bit less, at least. XD

We also might see a power reduction on the new chip as we did with X299 compared to X99. I don't know yet, but I'd imagine there are more changes there than simple feature changes, of which I expect very little.
 
Just like z270 wasn't the same as z170 because they added optane memory support.

Intel has moved new processes and new architectures onto the same chipset before. This is neither and they couldn't figure out how to make it work? I don't believe that for a second.
 
where can i find more info about z390. why would they release 2 motherboard like this in less than a year. what are the difference?
 
There is no doubt in my mind that Intel probably could keep its CPUs on the same socket for longer if it wished to do so, but there are some points you guys are ignoring. AMD's been constrained by aging sockets and put out worse products because of it. Bulldozer using AM3 wouldn't allow for full support of all the C-states that the CPU supported when used on older platforms. Then you end up with motherboards that can't handle the upper echelon CPUs because their VRM's can't handle the higher power requirements of those CPUs. You had 95w only capable motherboards that wouldn't support CPUs with a 130watt TDP. AMD wasn't able to switch to newer memory technologies as quickly, or as easily as Intel did. Again, constraining a CPU to an older platform may provide for a potentially long upgrade path, but not always a pleasant one. Socket longevity created BIOS and microcode issues with various motherboards. Intel does away with all of that crap by keeping its CPU support on a given socket and chipset limited. It also limits what can be done with the newer CPU beause of those constraints.

When you are trying to improve your CPUs on a performance per watt basis, part of doing that successfully is reducing the power requirements of the platform, or optimizing that platform to the CPU. AMD doesn't do that and ends up constrained by the electrical requirements of platforms that are dated to put it nicely.

I'm not saying that Intel couldn't put Coffee Lake on existing LGA 1151 based Z270 motherboards. Although, they have done it with refresh CPUs in the past which leads me to believe there is some sort of technical reason behind it. That said, the 300 series chipsets aren't the same as Z270 as others have claimed here. It looks like it will have an integrated audio DSP and support for a new C state, (C8). It's not a huge change but Z3xx doesn't appear to be the same exact chipset as Z270.
That would all be nice and reasonable if any of the generation/socket changes brought anything but miniscule improvements from one to the next. For such small increments, every second change would have sufficed.
 
where can i find more info about z390. why would they release 2 motherboard like this in less than a year. what are the difference?

Differences Listed here

Intel-Coffee-Lake-3-1000x516.jpg
 
Everyone bragged about how well Bloomfield, Sandy, Ivy, etc held up for 5+ years but now suddenly everyone wants to upgrade their Ryzen processor within 2 years. Maybe for workstation folks who want to move from 8C Ryzen chips to maybe 10C or 12C down the road? But for everyone else I just can't see it. I assumed Ryzen would get attention for its value proposition but the socket debate seems to be taking the crown. It's not even something I ever thought about before. My last CPU upgrade was after 7 years, this one will make 4 years. Both of which are longer than AM4's lifespan. I've never replaced a CPU in less than 4 years.

If you're the kind of person who needs to upgrade their CPU that often, or can afford to, then it seems like Intel would ultimately be a better choice since they are higher performing chips... right? Otherwise why else would you need that kind of up-to-date performance & core counts. Bizarre.

I suppose it's the easiest target in the AMD vs Intel debate, which is why people are hammering it so hard.

I think you missed the point, people that have Ryzen motherboards will have the option if they want to, not unlikely some updated motherboards will release with the next version of Ryzen as well. People have not felt a need to upgrade over the last 5 years due to meager IPC gains from Intel and well AMD was stuck with Bulldozer. I think most people expect Zen+ to be faster, not just more cores. It just gives a Ryzen owner options down the road is all. I used to upgrade every 2 years but my 8350 I held on to the longest and now that is more just a reality of how hard it is to make a faster processor that makes you feel the need to upgrade. So in the end it's just nice to have a option and I expect the 7nm Zen to be a nice upgrade. Intel has the best single threaded performance but AMD has the better multi threaded performance, it really depends on what you need your PC for. It's all a matter of needs.
 
Back
Top