Today is the day that Intel officially launched its new high-performance HEDT platform for enthusiasts. Two architectures join forces for the first time during the launching, Kaby Lake-X and Skylake-X, besides unveiling the Core i9 lineup, a new Intel X299 chipset and the new LGA-2066 socket with a 55 pin increase in contrast to the the previous LGA 2011.
Not every Skylake-X processor announced will arrive right away. Actually, the company has postponed the most interesting models’ arrival to the domestic market: Core i9 chips with 12, 14, 16 and even 18 cores and prices from $1,200 to $2,000.
In fact, the models that Intel launched today are an improvement on the models already existing in its Broadwell-X lineup, with models ranging from 4 to 10 cores. The first i9 to hit the market will have precisely 10 cores, the 10-core Core i9-7900X with 20 threads and a speed of up to 4.5 GHz.
We just said that this is the first time that two architectures coexist on a HEDT processors launching, and they do it in quite a peculiar way. If we review what Intel’s strategy has been during recent years for this platform for enthusiasts (refining the previous architecture for high-performance models), it was now Skylake-X’s turn, which actually is the architecture offering the most interesting models with more cores, from 6 to 18 cores.
The 7th-gen Kaby Lake’s strange presence comes with the most modest 4-core models, such as the Core i7-7740X, and even a Core i5 like the Core i5-7640X. It seems that Intel wanted to take advantage of its latest HEDT architecture for the first time, although they do not really offer anything new regarding the Kaby Lake architecture found on the LGA 1151.
However, one of the new features that we do find on Skylake-X is a new improvement for Turbo Boost, Turbo Boost 3.0.
The current Turbo Boost 2.0 technology gives us more performance by increasing the MHz of different cores at different speeds depending on their workload and speeds, therefore having a normal working speed, a boosted speed for all cores and a boosted speed for individual cores is typical.
With Turbo Boost 3.0, the processor is capable of detecting the two cores that are working the most in order to improve the performance on mono-core or dual-core tasks, thus increasing its speed above the maximum offered by Turbo Boost 2.0. This technology works hand in hand with Windows 10’s Anniversary Update or even with 2017 iterations of Linux kernels. Also, Intel will release some drivers to add support on PCs running earlier versions of Windows.
In addition to Turbo Boost 3.0, Intel has added other improvements such as SpeedShift in order to be more efficient at quickly activating sleeping cores when needed. The brand also adds the new AVX512 instructions focused on vectorization and performance improvements in complex operations.
Even if the HEDT platform has always been characterized by being the one delivering the most PCI-e lanes, Intel cut back on this aspect in the latest generations of its most ‘modest’ models, and it is happening again this time. Although the company has not unveiled its four most powerful models’ PCI-e lanes yet, the Core i7-7900X has 44 PCI-e lanes while the lesser models are reduced to 28 PCI-e lanes and the Kaby Lake-X models stay on absurd 16 PCI-e lanes, which are typical on the conventional LGA 115X platform, but not quite so on HEDT platforms, for example.
In fact, given Kaby Lake-X specifications, it seems that these processors do not have much reason to exist on a platform that has always been a demonstration of Intel’s power, either because of its high core count, its large PCI-e lanes number or its multi-channel memory, which are requirements that Kaby Lake-X does not meet at all by only offering 4 cores, 16 lanes and dual-channel memory.
Kaby Lake-X processors do seem to perform well when overclocked, and we hope for Skylake-X to do it too. There are already 8 speed world records and 23 first places on different overclocking competitions and rankings before its official launching. We actually already showed you how 7.5 GHz speeds were achieved at Computex.
As we were saying, the Skylake-X and Kaby Lake-X Intel cores will debut a new chipset named Intel X299.
Intel’s X299 chipset will be an adaptation of the technologies already featured by the brand on its Intel Z270, going from its ‘mainstream’ platform to the HEDT platform. We have Intel’s Optane Memory support to use an M.2 module as a cache drive, for example.
It features 24 PCI-e lanes, up to 8 SATA 3.0 ports and 10 USB 3.0 ports. Its controller allows you to connect up to 3 PCI Express 3.0 x4 for SSDs through Intel’s Rapid Storage Technology and has guaranteed support for the I219 LAN chipset.
The X299 is the first chipset debuting the new Socket R4, which is a socket that we will typically recognize as LGA 2066, where the entire Skylake-X and Kaby Lake-X lineup, from the most modest 4-core models to the more capable 18-core/36 threads ones, will coexist.
It is remarkable for Intel to be capable of building processors with so many cores using only 2066 pins, but we cannot forget that AMD will use a socket virtually doubling that pin count (4096) with its Threadripper lineup.
Certainly, the Core i9’s arrival is Intel’s answer to the great success that the RYZEN processors have had while also overshadowing its main rival’s Threadripper platform. Competition undoubtedly is a necessary condition in the market: a few years ago it was bold to think that we could have domestic processors with more than 10 cores, but we have gone from that amount to 18 cores in only two generations.
We will begin seeing Skylake-X and Kaby Lake-X’s real performance in the next few days. It all suggests that we will see a performance evolution linked to the new architectures, at least in the models featuring 4 to 10 cores. However, we will see the whole potential that Intel had up its sleeve the moment the 12, 14, 16-core models and the 18-core behemoth known as Core i9-7980XE are in our hands.
Read More:
Best Thermal paste for CPU