Teams from Intel and AMD have teamed up to crack down on Nvidia's mobile GPUs a little, and promise us thin and powerful gaming laptops. That is, the new Intel Kaby Lake G processors, enhanced by AMD Vega M graphics, can outperform GTX 1060 Max-Q cards while consuming less power. Sounds impressive, doesn't it?
As the January CES (Consumer Electronics Show) showed, the explosive event is about to happen this year; just at the start of Nvidia's open press conference - that big tech show - Intel announced its plans to break Nvidia's hegemony in the mainstream mobile gaming market.
Information for thought
Intel Kaby Lake G Release Dates
Machines equipped with the new Intel processors with Radeon graphics may arrive at the end of March. Hades Canyon Intel NUC minicomputers will ship at the end of March.
Specifications Intel Kaby Lake G
Kaby Lake G chips will be available with two main Vega M graphics options: the first with 20 compute units and 1280 GCN cores, and the second with 24 compute units and 1536 GCN cores. Both options provide 4 GB of HBM2 memory. All CPU components, including the Core i5, will be quad-core and eight-threaded.
Intel Kaby Lake G architecture
The G-series chips use CPUs with a relatively old 14nm Kaby Lake architecture, equipped with a modified Radeon Vega graphics chip connected via PCIe 3.0. The Vega M chip connects to the HBM2 memory via an internal Intel EMIB connection.
Intel Kaby Lake G Performance
Intel promises better gaming performance than Nvidia cards with both Vega M GH and Vega M GL graphics options, with chips with 24 compute units outperforming the GTX 1060 Max-Q by 10% in gaming tests, and chips with 20 computing units in some tests outperform the GTX 1050 by 40%.
The new Kaby Lake G processors promise mainstream gaming laptops that won't require additional heavy and hot discrete Nvidia or AMD graphics cards. Space savings alone make it possible to build laptops with larger batteries, more efficient and quieter fans, or simply smaller, lower power consumption gaming laptops.
The arrival of a mixed-chip with Intel Core CPU and Radeon Vega graphics shows just how much both companies want to push Nvidia out of the lucrative gaming laptop market. The gaming laptop market has grown by a total of 42% over the past three years, and this is in a world where Apple is trying to prove to you that the computer is dead and everyone else is saying that no one is buying desktop PCs anymore.
Despite a bitter relationship in the past, AMD and Intel reached a compromise on a number of contradictions - a purely monetary interest can be a good intermediary - because, as any expert knows, total war the enemy of my enemy is my friend. Or a custom integrated graphics provider.
Intel Kaby Lake G Release Dates
After a pre-announcement in January 2018 (before CES), we didn't expect to see laptops spinning new Intel/AMD hybrid chips until this spring. All in all, we thought the end of March was a very optimistic release date for any machines that could use Kaby Lake G/Vega M processors.
Intel has its own Hades Canyon NUC mini PC equipped with Vega M GH graphics that they are going to launch at the end of March, and we doubt that there are many laptop manufacturers that could beat Intel in terms of blanks with Vega M graphics. Although we know that Dell and HP definitely plan to release systems with new chips.
When we can finally get our hands on live laptops with Vega M GH graphics depends on specific manufacturers. Intel only talks about 100W of power for a NUC desktop mini with certainty, but we're almost hopeless to see all 1,536 GCN cores running in a compact gaming laptop capable of 1080p and 60fps.
Specifications Intel Kaby Lake G
This picture shows part of AMD's new Intel processor - quite an interesting piece. As you can imagine, talking about CPU components is boring - they use the terribly dull Kaby Lake architecture with a 14-nm process technology. This probably requires the right mood + knowledge of the inner workings, but I'm getting more and more tired of Intel's attempts to present the same architecture as something new in each release.
This means still the same four cores and eight threads across the board, without any of the mind blowing six core solutions that will delight the mobile market when they eventually release the series. Intel Coffee Lake-H in about a year.
However, of some interest is the Intel Core i5 chip with HyperThreading support and its own eight threads. This sets it apart from most Core i5 processors, and the only difference between it and the Core i7 is that it has a slightly lower clock speed and less total cache.
But, as I said, we are now looking at a really interesting Vega M graphics chip, which is offered in two different options: Vega M GH and Vega M GL, which means respectively high (Vega M Graphics High) and low (Vega M Graphics Low) graphics level.
The top-level graphics component Vega M GH in the G series is used only in chips with Core i7 and has a full set of 24 computing units (CU, Compute Unit). Each CU includes 64 GCN cores for a total of 1536 GPU cores. The frequencies of this GPU - both base and Turbo - are naturally much slower than comparable desktop Vega GPUs, but still reaching 1200 MHz is a very respectable result for a low-power chip that delivers this 100W TDP.
The Vega M GL processors include 20 CUs for a total of 1280 GCN cores. By comparison, that's 256 more cores than the RX 560 Polaris GPU. Since these chips deliver 65W TDP, their clock speeds will naturally be lower - in Turbo mode they only go to 1GHz.
It also appears from the specs that the GL chips that offer 32 pixels per clock have half as many ROPs as the GH chips that offer 64 pixels per clock. This indicator has highest value when it comes to post-processing and anti-aliasing, these settings may need to be lowered a bit if you're playing on a machine with a Vega M GL GPU.
In terms of memory, all G-series chips have 4 GB of HBM2 (High-Bandwidth Memory) memory that is connected directly to the GPU.
There is also one unlocked chip in the G series - Core i7 8809G, which recently appeared on the list of unlocked Intel processors, so there is nothing surprising here.
This means that with the Core i7 8809G, you, the lucky ones, will be able to use both overclocking applications - AMD's WattMan and Intel's XTU. And since the entire chip is unlocked, you get access to advanced settings for the CPU, GPU, and HBM2 memory. However, the other four G-series processors are completely blocked. Perhaps this suggests that the 8809G will remain a transition chip for desktop mini PCs such as the NUC Hades Canyon and will not go into G-series laptops with graphics high level Vega MGH.
The two chips, the i7 8809G and 8709G, are designed for the Hades Canyon NUC minicomputers, which John Deatherage, director of marketing for Intel NUC, called "Intel's Virtual Reality Machine" in a recent briefing. Now you understand why these computers were named Hades Canyon (Canyon of Hades), since their marketing director bears the name DEATHeRAGE, conducive to striving for the underworld of shadows ...
These will be surprisingly powerful compact machines, but claiming they can handle all the graphics requirements of VR gaming would be a bit of an overstatement. I understand that the NUC's GPU requirements are somewhat lower, but I think you would have to work hard to run Fallout 4 VR on the NUC in accordance with all the requirements for the gameplay.
Intel Kaby Lake G architecture
The fundamentals of the architecture of the new Kaby Lake G chips with Vega M graphics are already well known, with the exception of the complexities associated with the built-in EMIB (Embedded Multi-die Interconnect Bridge).
The architecture of the Kaby Lake CPU is over a year old - last January we presented it in the results of our tests. In addition, it is almost identical to the 14nm Skylake architecture that came out in 2015. But, as I said, this is in the order of things ...
In truth, the architecture of the AMD Vega GPU since its release last year has also become quite clear. Its key features are RPM (Rapid Packed Math) technology and HBCC (High Bandwidth Cache Controller). RPM essentially allows the GPU to execute two math instructions in the time of one, albeit with a slight loss in precision. But in games, this is not a problem, since there is no need for 32-bit precision calculations, unlike professional data processing.
The HBCC component allows the GPU to use part of the system memory as an extended frame buffer, which can come in handy when you only have 4 GB of video memory in a Vega M processor. This high-speed memory controller comes in handy when 4GB of HBM2 memory isn't enough. The presence of a 1024-bit memory bus implies a high bandwidth: 205 and 179 GB / s for the GH and GL chips, respectively.
With the Vega GPU, you also get access to all the latest AMD software updates. Last update AMD Adrenalin is the best driver they have released in the foreseeable past. For this type of mobile chip, Radeon Chill technology is excellent, which allows you to minimize energy costs to the utmost, and therefore save battery power during the game. You can also use FreeSync and FreeSync 2 technologies.
But perhaps the most interesting thing about this development is how Intel put it all together. They ordered a specially modified Vega GPU entirely from AMD, but used their own EMIB circuitry to connect the HBM2 to it. The EMIB method, which Intel introduced last year, allows different architectures and chips to be interconnected using a high-bandwidth bridge.
However, they did not use EMIB technology to connect Vega GPUs to Intel Core CPUs. This connection is very traditional way- using eight PCIe 3.0 lanes (PCIe 3.0 8x), while the other eight lanes are left for connecting to the CPU of a PCIe-based drive.
This is exactly the point that AMD could have done better than Intel, if you remember their own version of the integrated graphics in Ryzen mobile APUs. AMD's use of its own internal Infinity Fabric bus to connect the CPU and GPU in a single chip should be considered a better technical solution than the Intel Vega M layout, which is still essentially a simple combination of discrete GPU and CPU chips, rather than a highly efficient single chip. Will AMD release anything bigger than Ryzen mobile processors on its own? Most likely not - neither with the number of GCN cores that the modified Vega M GPU boasts, nor with the HBM2 video memory.
But Intel would likely defend its software-based dynamic power distribution scheme, drawing attention to the difference in performance between the two different approaches - team red and team blue - to using Vega graphics in mobile forms. Intel claims that Dynamic Tuning technology is almost 20% more efficient.
Vega also includes a per-CU power delivery system that allows the GPU to shut down entire clusters of GCN cores if they are not currently in use. And since the G-series is powered by Kaby Lake-H mobile components, you'll also get Intel HD graphics for those times when you don't need the high performance of Radeon graphics and are fine with a good midrange. Although I think Intel is exaggerating a bit when they say that the G series comes with “two amazing graphics subsystems.”
Intel Kaby Lake G performance
We will have to characterize the performance of the Kaby Lake G chips according to Intel, since the real machines that could show us new processors have not yet arrived on our test benches. We hope to have a wider selection of AMD Ryzen Mobile laptops available for benchmarking by then.
And, who knows, perhaps Nvidia will also release components for laptops based on the Volta architecture at the end of March. Yes, I doubt myself...
However, Intel's benchmark scores show that the top G-series Vega M notebook components are able to outperform the GTX 1060 Max-Q by an average of 10%, delivering 60 fps at 1080p at high settings. This is really impressive, even considering the fact that Max-Q Design chips are generally about 10% slower than standard Nvidia mobile graphics cards. So the Vega M GH graphics are potentially on par with the performance levels we currently see in gaming laptops starting at $1,500.
Now imagine how much laptops with Kaby Lake G processors will cost ...
Whether this performance will be enough for the Hades Canyon NUC with Vega M GH graphics to really qualify for true VR gaming remains to be seen. Although they call it a virtual reality machine, you will probably have to work with the NUC to get a fairly smooth gameplay experience in VR games with decent features - but not to forget about lunch and / or self-esteem .
The Vega GL chip was even more successful when compared to the corresponding component from Nvidia: Intel tests show that its performance exceeds that of a mobile chip Nvidia GTX 1050 by 30-40%. It is clear that Intel is showing results corresponding to the most optimistic scenario, but they are still impressive.
The results of the comparison with the GTX 1050 Ti were not presented, but it is known that the TDP of the Vega M GL GPU - 65 W - is almost the same as the total TDP (GPU + CPU), which is understandable. With Vega M GL graphics, you're unlikely to get 60fps at 1080p on high settings, but even hitting 40fps is a decent result. These are average figures, but it will be equally interesting to note the minimum frame rate and frame rendering time for both Vega M G-series chips.
The end of Intel's tick-tock manufacturing rhythm means that Kaby Lake has become the third architecture based on a 14nm processor. Starting with Broadwell (5th generation, tick), the manufacturer introduced a new microarchitecture Skylake (6th generation, "tock"), which was optimized in the 7th generation. Energy efficiency improvements and higher frequencies have been achieved through less stressful transistor layouts. Intel has launched a wide range of new Kaby Lake processors, ranging from the 15W and 28W mobile KBL-U and 45W KBL-H to KBL-S workstation models rated at 35-91W. There are also 3 overclockable options, including i3.
Lake Kabi
The first official launch of Kaby Lake took place in September 2016 and included 6 mobile processors designed for installation in premium laptops and mini PCs. They performed well, and at the beginning of 2017, Intel introduced more than 25 new models. The main feature of Kaby Lake processors is support for Optane Memory and 200-series chipsets. In addition, Gen9 graphics have been updated with Main10 and other low power video playback systems, and circuitry has been corrected to improve the voltage frequency curve.
Overview of Kaby Lake processors
Intel defines its product lines in Y, U, H, and S segments. Recent changes to the naming scheme have made it difficult to tell which segment a chip belongs to without knowing the TDP or core schematics.
Using the Kaby Lake Pentium, Core m3, Core i5/i7, and Core i5/i7 vPro nomenclature, the Y-series are hyper-threaded 2- and 4-core processors with a TDP of 4.5W that are targeted at small and light mobile PC. Such low power consumption is achieved thanks to the ultra-low base frequency. This allows you to install batteries with a smaller capacity, providing a small weight and long time offline work.
The U series consumes 28W and 15W, has 2 hyper-threaded cores but with a much higher clock speed. Includes Kaby Lake Pentium, Celeron, Core i3/i7 processors. They are often cheaper than the Y range because they are not limited by stringent voltage and frequency requirements and find use in premium gaming laptops. Some processors are equipped with an additional 64 or 128 MB eDRAM chip, which serves as a DRAM buffer with main memory and affects graphics speed.
The H-series chips have a rated power consumption of 45W and deliver the ultimate in mobile performance. Intel markets them under the VR Ready brand, which indicates their use in virtual reality systems. Available in various combinations of components and performance.
The S series is designed for desktop PCs. Nothing remarkable. There were 3 versions of the Core i7 released with 4 cores of the Kaby Lake processor and hyper-threading, one of which allows overclocking, and the other is low power. There are also several 4-core i5s in similar modifications and 2-core i3 chips.
The new KBL-S lineup features an overclockable Core i3-7350K, a hyper-threaded 60W dual-core processor with a base clock of 4.2GHz (no turbo) and a configurable multiplier. This was in response to the requests of enthusiasts, who thus achieve CPU performance corresponding to higher-end devices.
Speed Shift v2
One of Skylake's new features was the Speed Shift feature. With the correct driver present, the system may refuse to control the processor's turbo mode in favor of the processor itself. Using an internal metric collection, combined with access to system sensors, the CPU can adjust the frequency with greater accuracy and faster than the OS. The purpose of Speed Shift is to allow the system to respond faster to performance requests (such as interacting with a touch screen or browsing the web), reduce latency, and improve the user experience. Therefore, when the operating system is constrained by predefined P-state parameters, a Speed Shift-enabled processor with the right driver is able to change CPU frequency multipliers almost continuously over a wide range of values.
The first iteration of Speed Shift reduced the peak frequency gain time from 100ms to 30ms. The only limitation was the driver, which is now included with Windows 10 and shipped by default.
With the advent of the new architecture, the hardware control of Speed Shift has improved. Intel hasn't changed the name of the technology, but the improvements have been significant. The driver has not changed, so it works with all modifications of Speed Shift, but the processor can now reach its maximum frequency in 10-15ms, instead of 30.
Optane memory
One of the goals of the memory industry is to create something as fast as DRAM, but more durable, so that data can be retained even when there is no power. DRAM uses energy to update data, but is the main source of software data movement. Much of software acceleration depends on the speed of memory access or the ability to have data closer to the core when needed, so having large, close, non-volatile memory can increase performance and reduce power consumption. Most of the decade was spent on its creation. Intel (and Micron) officially announced their solution, 3D XPoint, a year ago, but it hasn't been officially released yet.
media opportunities
Although in terms of functionality, Intel Kaby Lake does not differ much from Skylake, there are clear improvements in the graphics. As with the CPU cores, the 14nm+ process allowed for higher frequencies and improved GPU performance, but perhaps more impressive changes are the upgraded media capabilities. The core architecture of the Gen9 GPU has not changed, but Intel has revised the video processing units to add functionality and improve efficiency.
4K hardware acceleration
The main difference in the Kaby Lake-U / Y media engine is the presence of full hardware acceleration for encoding and decoding 4K videos in the HEVC Main10 format. This is in contrast to Skylake, which supports 4k p30, but does so using a hybrid process that shares the load between the CPU, media processors, and GPU shader cores. As a result, Kaby Lake not only handles more HEVC profiles, but consumes only a fraction of the power at a much higher throughput. Also in the new architecture, 8-bit encoding and 8/10-bit decoding of the VP9 codec from Google were implemented. Skylake offered hybrid codec decoding, which did not provide sufficient power efficiency. The new HEVC Main10 and VP9 hardware acceleration scheme is part of the MFX block. The video quality engine received support for HDR and Wide Color Gamut.
According to Intel, Kaby Lake U/Y is capable of handling up to 6 4K30 AVC and HEVC codecs simultaneously. Support for HEVC decoding is rated at 4K60 up to 120 Mbps, which is necessary for playing premium content and UHD Blu-ray. Thanks to process improvements, even 4.5W Y chips are able to process HEVC 4Kr30 in real time. Thus, in the U and Y series, one of the main complaints about Skylake was resolved: the lack of hardware accelerated 4Kp60 HEVC Main10 decoding. There are other improvements that provide a more satisfying multimedia experience for consumers.
Connectivity
Kaby Lake U/Y processor graphics flow is the same as Skylake. This means that the iGPU serves up to 3 displays at the same time.
One disappointing aspect of Skylake that wasn't addressed in Kaby Lake-U/Y is the lack of a native HDMI 2.0 port with HDCP 2.2 support. Intel is in favor of adding LSPCon to DP 1.2. This approach has been used on several motherboards and even mini PCs such as the Skull Canyon NUC (NUC6i7KYK) and the ASRock Beebox-S.
Chipsets
The new PCH controller hubs are paired with LGA1151 sockets and thus support both Skylake and Kaby Lake. 100 series chips such as the Z170 are also compatible with new processors after a BIOS update.
Today is pretty predictable. The Z-series is focused on multi-graphics chips and overclocking, H is notable for the absence of the latter, Q is intended for platforms with vPro support and B is focused on cheaper solutions.
There are also 3 mobile chipsets available with similar differences, including a Xeon kit in the CM238 that allows the use of the new E3-1500 v6 processors.
Compatible Boards
Motherboards for Kaby Lake processors - ASUS Maximus IX Code, GIGABYTE Z270X, Supermicro С7Z270-CG, ASRock Z270, MSI Z270, ECS Z270H4-І. They have new controllers, including the USB 3.1 10Gb/s ASMedia ASM2142, which uses two PCIe 3.0 lanes to support up to 2 ports. Previously, only 1 PCIe 3.0 slot was used for this.
The Realtek ALC1220 audio controller has also been updated: there is a 120 dBA output and a 113 dBA input. This should provide the best measurable quality. The network connection is still handled by the Intel I219-V gigabit Ethernet controller. The big change here should be the introduction of the multi-gigabit Aquantia 5G/2.5G AQC107. New is the 10Gbps USB 3.1 interface on the front of the MSI Z270 Gaming M7. It is currently activated via ASM2142 using two PCIe lanes to provide one USB 3.1.
Technically, all motherboards equipped with Kaby Lake should be able to support Optane Memory. LED backlighting also plays a big role in the 200-series motherboards: only a few models in each price category are deprived of it.
Performance
As expected, there is no performance gain. Based on user feedback, the 3GHz Kaby Lake i7-7700K performs similarly to the 3GHz Core i7-6700K (with hyperthreading disabled). The only difference is in memory support. Whereas Skylake is DDR4-2133 compatible, Kaby Lake is DDR4-2400 compatible, however this does not significantly affect almost all benchmarks.
Power consumption
One of the main advantages of the Kaby Lake processor is the same frequency with less power or more with the same power compared to Skylake. The i7-7700K supports 4.5GHz turbo mode with 91W of thermal power. All tested Kaby Lake processors, even with manual overclocking, have close to calculated consumption, although usually the CPU supplier significantly overestimates the voltage required for stable operation of the chip.
Overclocking
According to user feedback, their perception of the increase in clock speed in Kaby Lake has changed due to new feature AVX Offset found in the BIOS of every Z270 motherboard. AVX instructions are known to be detrimental to overclocking, reducing stability and making it difficult to move code without AVX. The user can now apply an offset (eg -10x) which will decrease the multiplier when the AVX command is encountered. This means that when overclocking a Kaby Lake CPU to 4.8GHz with 8x AVX bias, the AVX command will run at 4.0GHz, generating less heat and keeping the system stable.
According to users, the 4.8GHz AVX frequency is easily achievable even at a reasonable voltage. The i7-7700K reaches 4.9GHz with an AVX offset of -10, while the i5-7600K reaches 5.0GHz even with AVX enabled.
By and large, overclocking the i7-7700K from 4.2 to 4.8 GHz does not provide a practical advantage. A difference of 600 MHz corresponds to a 13-14% increase in performance, which is not much. However, given the voltage profile of the chips, 4.5GHz provides good temperatures and voltages, still outperforming the i7-4790K or i7-6600K.
Test results
Based on user feedback, the Kaby Lake CPU comparison confirms that the Core i7-7700K wins in almost every test (except for a few where the i7-5775C still outperforms due to 128MB eDRAM).
The Core i5-7600K performs pretty much the same except for low-threaded scenarios (such as ray tracing), but the processor certainly doesn't fall short in day-to-day tasks. The Core i5-7600K, due to the lack of IPC growth, is essentially the base i5-6600K, aside from a few extra megahertz. The processor overclocks well - its temperature is much better than the i7-7700K, but it is nothing more unusual.
The elephant in the china shop, however, is the Core i3-7350K. At $159, it's only $11 from the Core i5-7400, which costs $170 but has 2 two full cores, albeit at a lower frequency (3GHz vs. 4.2GHz).
Is the new Intel architecture a new milestone?
For the most part, Kaby Lake doesn't offer much change. Support for Optane memory is a plus, but otherwise it's just a shift in the power/efficiency curve. Power consumed at 3.0 GHz last year is now 3.3 GHz, which means saving time spent doing work or saving electricity. Speed Shift v2 is a really nice feature, but limited to Windows 10 users. Of more interest is the set of new controllers (ALC1220, E2500, Aquantia). The optimization architecture does not cause admiring surprise, but provides a 10% increase in efficiency.
Kaby Lake (pronounced Kaby lake) is the code name for the seventh generation of Intel Core processors.
The 14nm Intel Kaby Lake processors differ from the Skylake processors in a minor change in the Core microarchitecture in line with Intel's microprocessor development strategy.
Peculiarities:
- 14nm process technology.
- Design LGA 1151.
- Support for PCI Express 3.0.
- Thunderbolt 3 support.
- Support for USB 3.1, unlike Skylake, which requires additional controllers on the motherboard for USB 3.1 ports to work.
The Kaby Lake family is divided into three segments: K - unlocked models, S - "standard" models (no suffix in the name), T - chips with a lower TDP.
Simultaneously with these processors, the 200 series chipsets will debut, which will include the following models: Z270, H270, Q270, Q250 and B250.
Along with the release of desktop processors "family" Kaby Lake, Intel is preparing to update the CPU line for entry-level professional systems for the LGA1151 platform.
The Xeon E3-1200 v6 processors will replace the existing Xeon E3-1200 v5 processors based on the Skylake architecture.
Unfortunately, the new line, like the current one, will only be compatible with Intel chipsets C232 and C236, and when you install such CPUs in a desktop motherboard designed for "regular" Kaby Lake-S processors, they simply refuse to work.
The frequency increase relative to the v5 line is 100-200 MHz, depending on the specific model, and the integrated graphics core in the Xeon E3-12x5 processors has not changed at all.
With the release of Kaby Lake processors, Intel has improved its 14nm process and achieved a good level of chip yield.
Intel has adjusted the profile of its 3D tri-gate transistors, and thanks to this, 14nm semiconductor chips were able to get better frequency potential.
It would be fair to call Kaby Lake Skylake Refresh, because then it would be quite clear that no improvements at the microarchitecture level should be expected.
In Kaby Lake, compared to Skylake, there is not even the usual three to five percent increase in performance.
At clock speeds equal to Skylake, the new processors deliver absolutely identical performance, and all their advantage is explained only by the operating frequencies increased by 200-300 MHz.
However, the previously missing full support for HEVC and VP9 formats with both 8- and 10-bit color was added to the graphics core of the processor, which is responsible for hardware encoding and decoding of video content, which should positively affect the autonomy of mobile devices when playing video.
Enthusiasts will be pleased with the increase in overclocking potential, resulting in overclocking Core processors The i7-7700K and Core i7-7600K, as well as the inexpensive unlocked core i7-7350K that joined them, are capable of reaching 5 GHz with conventional air cooling.
What is it that attracts them in the first place?
Kaby Lake is the next generation of processors from Intel. We are currently using the SkyLake generation. At least most of us, if you are not in a hurry to buy an updated .
You will still see laptops for sale with previous generations of processors, like Broadwell and Haswell, but officially they are already in the past.
In this article, we have collected all the details you need to know about the upcoming revolution in the world of Intel Core Kaby Lake processors.
In pursuit!
- What's this? 7th generation Intel Core processors;
- When to expect? Laptops out now, PCs in Q1 2017;
- What is the price? Pricing is similar to modern Intel Skylake;
ProcessorsIntelKabyLake: Release Date
July 22 CEO Intel, Brian Krzanich, has confirmed that Kaby Lake chipsets have made their way from development fields to factory conveyor belts and then to PC manufacturers. In other words, Kaby Lake processors are officially on the doorstep.
This means we could expect some Kaby Lake (PCs) before the end of 2016. However, at the moment it is not known exactly which chipsets will come in the first wave.
Intel Kaby Lake includes Intel Core i3/i5/i7 desktop and laptop processors and new Core M.
Even after Intel's keynote at Intel's own Developer Forum in San Francisco, California, we don't know a release date for Intel's 7th generation desktop processors, but all signs point to CES in January, at least that's what they think some publications, and we agree with them.
At the same time, we have not experienced a lack of information leaks regarding the new Kaby Lake processors and the release date. Some tech publications like WCCFtech have unearthed documents that point to prices and specifications, while the folks at Tom's Hardware claim to have bought their own (possibly retail) Kaby Lake processor.
ProcessorsIntelKabyLake
In addition to the mobile series, 20 Kaby Lake processors are waiting for their users on sale. From the Pentium G3930 to the Core i7-7700K, there's almost a complete selection available with the latest generation.
The Kaby Lake Core i7-7700K is the flagship processor this time, unlocked for overclocking, as indicated by the "K" in the name. New series Kaby Lake continues to use the company's serial names: "7" indicates the Kaby Lake series of processors, as it is the seventh generation, and Skylake is the 6th generation with "6" in the number.
The Core i7-7700K is a 4-core hyper-threaded processor, and while early benchmark results (for March) promised us clock speeds between 3.6GHz and 4.2GHz (Turbo Boost), recent reports tease fans a lot more. fruitful 4.2GHz / 4.5GHz. Of course, actual results may vary.
The original leaks come from SiSoft's benchmark benchmark database, but unfortunately these data are significantly worse than the current generation i7-6700K. The positive side of the rumors promises us more reliable "boost" per core, 200 MHz / 500 MHz (Boost), respectively, compared to its predecessor.
The leaks also hint at a price tag of $350 ($22,000), which is very close to the cost we'd expect with an equivalent Skylake generation processor at release.
Next up is the Core i7-7500U, which went online alongside the i7-7700K. This is the kind of CPU we'd eventually expect to see in high-end ultrabooks. It's a relatively high performance chipset, but it still bears the "U" in its name, meaning it belongs to the ultra-low voltage family.
It has two cores, four threads and is clocked at 2.7GHz - 2.9GHz (Turbo). Some of you may turn up your nose at dual-core chipsets on laptops, but they play an important role.
On the mobile front, the previous generation Core M5 and M7 are now integrating the "Y" in the Core M family. These include the Core m3-7Y30, Core i5-7Y54, and Core i7-7Y75, which are used in leading laptops with fanless designs and convertible formats to addition to U-series processors.
The first laptopsIntelKabyLake
Where will we see these chipsets end up? Well, they currently appear on a short list of laptops, some of which have already made it through our reviews. The new chips are featured on the Razer Blade Stealth and HP Specter x360, along with ultrabooks, 2-in-1 hybrids, and traditional laptops, among a host of others.
If you're wondering why the latest MacBook Pro is still clinging to Skylake, the answer is simple: at the time of the laptop's release, the required Kaby Lake processor series didn't exist yet. Thankfully, DigiTimes reports that we'll be seeing high-end laptops with these chips at CES in January.
Some say that Apple may skip Kaby Lake altogether, but this seems unlikely as the next generation of Cannonlake is not expected until the second half of 2017. According to the schedule, the 12-inch MacBook should receive 7th generation Intel processors this spring.
ArchitectureIntel Kaby Lake
Cannonlake is likely to be much more exciting than the Caby Lake update. You see, Kaby Lake is very similar to the Skylake family. This is not what we expected from the successor to Skylake, but Intel has changed the strategy for developing its processors.
Since 2007, Intel has followed a "tick, tock" upgrade mode, where one generation leads to a smaller processor, and the next generation changes the architecture. The situation has changed this year. As of 2016, Intel is using "Process, Architecture, Optimization" as an approach, and KabyLake represents, frankly, not the most interesting stage.
It's still a 14nm processor that's broadly similar to Skylake, and desktop processor models will use the same LGA 1151 socket. If all goes well, Cannonlake promises to shrink processors to the long-promised 10nm in 2017.
And while there are probably some performance and overall efficiency improvements to come, we feel there is no need for Skylake processor owners to upgrade to the same level of KabyLake.
UpdateIntelKabyLake
There are a few different improvements specific to Kaby Lake, though. Fully integrated USB-C Gen 2 support comes first. Skylake is offering support now, but requires additional hardware. Soon the technology will become "native". Again, an interesting solution, but not necessary.
Gen 2 USB 3.1 provides 10Gbps of bandwidth instead of 5Gbps. Thunderbolt 3 support is also there. In the same vein comes support for HDCP 2.2. This is digital copy protection, a new version designed for certain 4K video standards. Ultra HD Blu-Ray is becoming key, though 4K Netflix video also requires Kaby Lake processors.
It's also true that Kaby Lake also offers integrated GPUs that are better suited for 4K video. Thanks to the new media engine based on the Gen9 graphics architecture, users will be able to edit 4K video in real time using nothing more than integrated graphics. In terms of video consumption, the new VP9 and HVEC 10-bit decoder will allow you to stream 4K video all day on a single charge.
Kaby Lake processors also officially support Windows 10 among operating systems Microsoft. This is another attempt by Microsoft to push those who linger on Windows 7 and other operating systems.
Apollo Lake: Poor RelativeKabyLake
Also worth considering are the Atom chipsets, which occupy the bottom of the series and will be used in very cheap Windows 10 laptops and tablets. Although they are not part of the Kaby Lake series, the latest "Apollo Lake" chips began to appear at the end of November, ASUS and HP are among the first to implement new processors.
They are also capable of accelerating 4K video playback thanks to the HEVC and VP9 codecs. This is partly due to the transition of Gen8 graphics to Gen9 graphics, just like the Skylake processors.
Kabylake-X: Best Last
If you're only interested in the mainstream Kaby Lake processors, the future doesn't look too bleak. They will go into production before being replaced by Cannonlake at the end of 2017. However, the outlook for serious high-end chips is far more confusing.
Now the latest high-performance Intel processors are part of the Broadwell-E series, although among the mainstream processors, Broadwell has become obsolete. Simply put, real high-end hardware will come later. We are talking about processors like the Core i7-6900K for 100,000 rubles.
The Kaby Lake alternative will not be called Kaby Lake-E, instead we are waiting for Kaby Lake-X, which is expected to launch in the second half of 2017, along with Skylake-X. That's right: two generations at the same time.
Intel Kaby Lake-X will be tentatively a 4-core processor, while Skylake-X will be a very puzzling 10-core processor.
What mere mortal laptop and desktop buyers should know about Kaby Lake, however: a) we'll see more machines using the new chipset sets very soon and b) unless you're in need of an upgrade right now, 2017 will bring Cannonlake with interesting improvements .
In short: we have the same Skylake chips, but with higher frequencies and an advanced video processing hardware engine. And yet some models are very interesting. In addition, there is an unshakable rule: it is better to assemble a computer from scratch on the most modern hardware possible.
Intel Core i3-7320
Briefly about the product: 2 cores but 4 threads, 4.1GHz, 4MB L3 cache, 51W TDP
Peculiarities: very high frequency in default - 4.1 GHz
Price:$149
Budget gaming computer with this processor: 35-40 000 rubles
Initially, this place in the selection was assigned to the Core i3-7350K. He is unique. As the song of the Kino group says: our hearts demand change! Indeed, since 2011, Intel has had two overclockable processors. One Core i5 and one Core i7 (there was also an anniversary Pentium G3258, but this is an exception that proves the rule). Such patterns are easy to recognize. They are the fastest, they are the most expensive, they have the letter "K" in the name. The wind of change blew precisely in 2017, with the release of the Core i3-7350K. For a long time, Intel has not released overclocking budget processors. Naturally, you will have to pay extra for overclocking capabilities. The chip costs $168, but it's still cheaper than the slowest Kaby Lake quad-core Core i5-7400 ($182).
The Core i3-7350K is fast and without any overclocking. Operates at a frequency of 4.2 GHz. It is quite possible to overclock it up to 4.8-5.0 GHz. Naturally, for this you will need to have a high-quality cooler in your arsenal. And in general, overclocking requires a more expensive motherboard based on the Z170/Z270 Express chipset. Read about what devices are needed for the seventh generation Core in this material. So saving is a moot point. As well as the possibility of overclocking. But 4.2 GHz out of the box is already serious. And the Core i3-7320 runs at 4.1 GHz. Only 100 MHz less, but we save $19 right away.
Intel Core i3-7320
Intel Core i5-7400
Briefly about the product: 4 cores, 3.0 (3.5) GHz, 6 MB L3 cache, 65 W TDP
Peculiarities: cheapest quad-core Kaby Lake
Price:$182
Gaming computer budget: 50-55 000 rubles
And Core i5 processors, as you know, have full four cores. And modern games love multithreading more and more. Perhaps the most obvious example is Battlefield 1. In it, any Core i5 is 100% loaded. But such a chip is still enough to build a gaming computer with a powerful graphics card, including the Radeon RX 480 and GeForce GTX 1060.
Do not forget about one tempting feature of the new Kaby Lake. The chips received not very fast HD 630 integrated graphics, but it has an advanced media block. As a result, all the forces of the processor can be "thrown" to ensure the operation of the video card, and the hardware blocks of the integrated GPU, for example, will ensure the operation of the OBS streaming program.
Intel Core i5-7400
Intel Core i7-7700
Briefly about the product: 4 cores but 8 threads, 3.6 (4.2) GHz, 8 MB L3 cache, 65 W TDP
Peculiarities: fastest processor with 65W TDP
Price:$303
Gaming computer budget: 60-75 000 rubles
The capabilities of the Core i7-7700 are studied in detail in the review. The most "soft" is that at a fairly low for desktop TDP processors(65 W in total) under load, all four cores of the chip operate at a frequency of 4 GHz. We get two things. Firstly, there is sense from eight streams, including in games. Secondly, high frequency. Helps with work and play. The Core i7-7700 will make great friends with a GeForce GTX 1070-level video card. A low level of typical heat dissipation will allow you to assemble a gaming computer of any complexity. Yes, even the size of a game console!
Intel Core i7-7700
Intel Core i7-7700K
Briefly about the product: 4 cores but 8 threads, 4.2 (4.5) GHz, 8 MB L3 cache, 91 W TDP
Peculiarities: accelerates up to 5 GHz. If you're lucky.
Price: $339
Gaming computer budget: 100 000 rubles
Intel's mainstream platform, which is what LGA1151 is, supports a maximum of quad-core Core i7 processors. Therefore, the Core i7-7700K differs from the Core i7-7700 only in frequency, the presence of an unlocked multiplier and, as a result, an increased TDP level. Overclocker model. With proper luck, it can be overclocked to 5 GHz using a good cooling system. The last time such overclocking agility was boasted by Sandy Bridge chips, released back in 2011. It is clear that any modern video card can be used with the Core i7-7700K. Or even two.