Coffee Lake Not Supported by Intel’s 200-Series Motherboards
by Ian Cutress on August 2, 2017 5:12 PM EST- Posted in
- CPUs
- Intel
- LGA1151
- 200-series
- Core 8th Gen
- Coffee Lake
- 300-Series
In a stunning bit of Twitter, a tweet from one of the leading motherboard manufacturers has stated that Coffee Lake, Intel’s 8th Generation Core processors, will not be supported on the current generation of 200-series motherboards.
No，Coffee Lake CPU is not compatibilble with 200 series motherboards.
— July 31, 2017
Information like this is usually kept under wraps until an Intel reveal, but it seems to have been mindlessly posted to Twitter on July 31st, an account that last tweeted on April 11th before this tweet occurred. This tweet has since been deleted.
At this point, due to the similar microarchitecture to Kaby Lake being used in Coffee Lake, most of the technology press were under the impression that the Coffee Lake processors would be compatible with LGA1151 socket motherboards, namely the 100-series and 200-series. With the above tweet essentially confirming that Coffee Lake will not be supported, it means that either the new CPUs will not be LGA1151, or that the motherboards will lock-out the processors by firmware, or the CPUs and sockets will use a different notching system to ensure the wrong processor cannot be put in the wrong board. It does mean however that 200-series users hoping to upgrade to a Coffee Lake processor (which early reports are suggesting might be up to six cores, but this has not been announced) will not be able to.
There are many potential reasons for the change if the socket is still LGA1151. The obvious one would be product segmentation on Intel’s part, which would stick in the craw for a number of the user base. The second one that it might actually be a physical requirement for the processor – if previously unused pins are required for power and/or control for different elements of the DVFS in the chip. This would depend on new features on the chip, which could extend to different power management, different graphics, or different IP blocks that require separate pin-out connections. Intel might also be using a different power system for voltage regulators, which might not be compatible with current 200-series motherboards.
At this point, nothing has been made official. The fact that this was stated on Twitter so far from any launch date that we know of is an interesting development.
*The name of the manufacturer has been removed by request after this news was published.
Related Reading
- The Intel Core i7-7700K (91W) Review
- The Intel Core i5-7600K (91W) Review: The More Amenable Mainstream Performer
- The Intel Core i3-7350K (60W) Review: Almost a Core i7-2600K
- The Intel Skylake-X Review: Core i9 7900X, i7 7820X and i7 7800X Tested
- The Intel Kaby Lake-X i7 7740X and i5 7640X Review
- The AMD Zen and Ryzen 7 Review: A Deep Dive on 1800X, 1700X and 1700
- The AMD Ryzen 5 1600X vs Core i5 Review: Twelve Threads vs Four at $250
- The AMD Ryzen 3 1300X and Ryzen 3 1200 CPU Review: Zen on a Budget
Source: Twitter
Facebook
Twitter
RSS
134 Comments
View All Comments
Kvaern1 - Thursday, August 3, 2017 - link
Speaking of unnecessary chipsets, can anyone explain to me why the Kaby Lake AVX downclock option coudn't be enabled on Skylake/z170 setups via a BIOS update?TEAMSWITCHER - Thursday, August 3, 2017 - link
Not very likely... When the list price of the 6-core Skylake-X Core i7-7800X is $389.Gothmoth - Wednesday, August 2, 2017 - link
not really a surprise when you know intel.they always thought "we have a loyal customer base that is not very bright" let´s milk them.
honestly, looking around the web today i see intel fanboys defending this decision from intel by saying coffee lake would need a new chipset to show it´s full potential.... ROTFL.
you have to love such customers as a company.......
Old_Fogie_Late_Bloomer - Wednesday, August 2, 2017 - link
Or, you know, it could be because Cannon Lake is adding native support for USB 3.1 Gen 2 and the old motherboards weren't designed to handle that.Intel is not a great company, but the need for a new motherboard shouldn't come as a surprise to anyone who has actually been paying attention.
Old_Fogie_Late_Bloomer - Wednesday, August 2, 2017 - link
^ Coffee* Lake. Frigging code names and no edit button...HStewart - Wednesday, August 2, 2017 - link
You could still be right even with "Cannon Lake" - "Coffee Lake" maybe use a transitional platform for 10nm "Cannon Lake" platform.Intel is not the only company that switches motherboard designed, new AMD chips don't support older platform.
Keep in mind also - there could more that one Coffee Lake chip - maybe would that can support the older motherboard but loose some of features plan for newer platform.
ddriver - Wednesday, August 2, 2017 - link
However the cases where a given intel chipset supports more than a single generation of processors are few and far in between. Intel's platform longevity is close to ZERO.AMD also changes platforms, but only every once in a while, and only when the change is a technical requirement. AMD platforms are designed for longevity, for example the current zen chipsets will support zen 2 and zen 3, and not just the first current zen, well into 2020 and beyond.
Intel in contrast, regardless of what this article might allege, chance chipsets simply in order to sell more chipsets, they'd rather shove a whole new chipset and integrated graphics most people don't want down their throats and pocket in the money.
Not that AMD wouldn't love to be in the position to do the same, but they don't have a monopoly to exploit, being the underdog they are simply forced to play nicer.
nevcairiel - Thursday, August 3, 2017 - link
Its a shame though, because AMDs chipset IO connectivity is extremely limited already, if they won't extend it for ~5 years or so, that will make the entire platform look terrible by the end of it.ddriver - Thursday, August 3, 2017 - link
How exactly is it limited?bill.rookard - Thursday, August 3, 2017 - link
Exactly. Looking at the block diagram, the Ryzen AM4 socket provides natively:4 x USB 3.1 Gen 1 / 16 x PCIe 3.0 (graphics) / 4 x PCIe 3.0 (M.2 @ x4 or 2 x M.2 @ x2 or SATA 3.0 x 2)
Connected to the chipset via PCIe x 4 equivalent to DMI 3.0 - chipset provides:
2 x USB 3.1 Gen 2/6 x USB 3.0/6 x USB 2.0
8 x PCIe 2.0
6 x SATA 3 (depending on the chipset - up to 6).
Meanwhile, LGA 1151 provides
16 PCIe 3.0 lanes (graphics) (yeah - that's it)
connected to the chipset via PCIe 3.0 x 4 to the Z270 which provides :
up to 24 PCIe 3.0 i/o lanes / Gigabit ethernet / 10 USB 3.0, 14 USB 2.0.
So - looking at the block set - from the point of (for example) a program importing a video on the fly, compressing it, and putting it out back to storage:
On the Ryzen CPU, if you have the storage attached to the CPU based SATA ports and/or USB 3.1 ports, the transfers are direct from the storage to the CPU across dedicated links, processed, then back out across the dedicated links. You could theoretically max this out if the storage is the USB 3.1 links and the CPU SATA or NVME drives. If you're using the chipset based SATA/USB 3.1 ports, then it follows the same path as Z270, with the chipset and the 4 PCIe 4.0 CPU links getting in the way.
Meanwhile, the LGA 1151 CPU has NO dedicated IO direct to the storage/USB connections. All storage and connectivity goes through the DMI (PCIe x4) links so all data coming in must be processed through the chipset across the DMI links which is where your bottleneck might show up.
That means in certain scenarios/workloads and depending on how you connect your storage, a Ryzen AM4 setup could be faster than the associated LGA 1151 CPU because of how some of the storage is connected directly to the CPU.