A surprise twist from Intel
Intel is going to double the core count on 15-watt processors starting next month.
Any expectations I had of a slower and less turbulent late summer and fall for the technology and hardware segments is getting shattered today with the beginning stages of Intel’s 8th Generation Core Processors. If you happen to think that this 8th generation is coming hot on the heels of the 7th generation that only just released to the consumer desktop market in January of this year, you’d be on the same page as me. If you are curious how Intel plans to balance Kaby Lake, Coffee Lake, and Cannon Lake, all releasing in similar time frames and still use terms like “generation,” then again, we are on the same page.
Today Intel launches the 15-watt version of its 8th Generation Core Processors, based on a refresh of the Kaby Lake CPU design. This not a new architecture nor is this is not a new process node, though Intel does talk about slight changes in design and manufacturing that make it possible. The U-series processors that make up the majority of the thin and light and 2-in-1 designs for consumers and businesses are getting a significant upgrade in performance with this release. The Core i7 and Core i5 processors being announced will all be quad-core, HyperThreaded designs, moving us away from the world of dual-core processors in the 7th generation. Doubling core and thread count, while remaining inside the 15-watt thermal envelope for designs, is an incredible move and will strengthen Intel’s claim to this very important and very profitable segment.
Let’s look at the specifications table first. After all, we’re all geeks here.
| Core i7-8650U | Core i7-8550U | Core i5-8350U | Core i5-8250U | Core i7-7600U | Core i7-7500U | |
|---|---|---|---|---|---|---|
| Architecture | Kaby Lake Refresh | Kaby Lake Refresh | Kaby Lake Refresh | Kaby Lake Refresh | Kaby Lake | Kaby Lake |
| Process Tech | 14nm+ | 14nm+ | 14nm+ | 14nm+ | 14nm+ | 14nm+ |
| Socket | BGA1356 | BGA1356 | BGA1356 | BGA1356 | BGA1356 | BGA1356 |
| Cores/Threads | 4/8 | 4/8 | 4/8 | 4/8 | 2/4 | 2/4 |
| Base Clock | 1.9 GHz | 1.8 GHz | 1.7 GHz | 1.6 GHz | 2.8 GHz | 2.7 GHz |
| Max Turbo Clock | 4.2 GHz | 4.0 GHz | 3.8 GHz | 3.6 GHz | 3.9 GHz | 3.5 GHz |
| Memory Tech | DDR4/LPDDR3 | DDR4/LPDDR3 | DDR4/LPDDR3 | DDR4/LPDDR3 | DDR4/LPDDR3 | DDR4/LPDDR3 |
| Memory Speeds | 2400/2133 | 2400/2133 | 2400/2133 | 2400/2133 | 2133/1866 | 2133/1866 |
| Cache (L4 Cache) | 8MB | 8MB | 6MB | 6MB | 4MB | 4MB |
| System Bus | DMI3 – 8.0 GT/s | DMI3 – 8.0 GT/s | DMI2 – 6.4 GT/s | DMI2 – 5.0 GT/s | DMI2 – 5.0 GT/s | DMI2 – 5.0 GT/s |
| Graphics | UHD Graphics 620 | UHD Graphics 620 | UHD Graphics 620 | UHD Graphics 620 | HD Graphics 620 | HD Graphics 620 |
| Max Graphics Clock | 1.15 GHz | 1.15 GHz | 1.1 GHz | 1.1 GHz | 1.15 GHz | 1.05 GHz |
| TDP | 15W | 15W | 15W | 15W | 15W | 15W |
| MSRP | $409 | $409 | $297 | $297 | $393 | $393 |
The only differences between the Core i7 and Core i5 designs will be in cache size (Core i5 has 6MB, Core i7 has 8MB) and the clock speeds of the processors. All of them feature four true Kaby Lake cores with HyperThreading enabled to support 8 simultaneous threads in a notebook. Dual channel memory capable of speeds of 2400 MHz in DDR4 and 2133 MHz in LPDDR3 remain. The integrated graphics portion offers the same performance as the 7th generation designs, though the branding has moved from Intel HD Graphics to Intel UHD Graphics. Because Ultra.
But take a gander at the clock speeds. The base clock on the four new CPUs range from 1.6 GHz to 1.9 GHz, with 100 MHz steps as you go up the SKU ladder. Those are low frequencies for modern processors, no doubt, but Intel has always been very conservative when it comes to setting specs for base frequency. This is the speed that Intel guarantees the processors will run at when the CPU is fully loaded using a 15-watt TDP cooling design. Keeping in mind that we moved from dual-core to quad-core processors, it makes sense that these base frequencies would drop. Intel doesn’t expect users in thin and light machines to utilize all 8 threads for very long, or very often, and instead focuses on shorter use cases for multi-threaded workloads (photo manipulation) that might run at 3.x GHz. If this period of time is short enough, the cooling solution will be able to “catch up” and keep the core within a reasonable range.
The single threaded Turbo Speeds increase by 200 MHz as we move up the SKUs, hitting as high as 4.2 GHz on the Core i7-8650U. That creates a 2.3 GHz Turbo Boost range on that part, leaving Intel a lot of room to manipulate the voltages and clock speeds in order to maintain maximum performance in that slight, 15w thermal envelope. It also means that single threaded performance should be faster than we have seen from 15-watt parts previously, as the 8650U peaks at 200 MHz faster than any 7th generation mobile processor.
Even the Core i5-8250U will offer exceptional single and multi-threaded performance, combining a quad-core design (with HyperThreading enabled) that can scale between 1.6 GHz and 3.4 GHz.
The significance of this move in the thin and light market is substantial. Intel will essentially be able to offer nearly double the performance of the previous generation platforms in peak theoretical capability, though with some performance write off due to the need to clock lower to stay within thermal limits. But those thermal limits will be less of a problem for shorter, burst workloads. We have often talked about short, single threaded “bursty” workloads like touch screen recognition in our mobile stories, and the impact that high Turbo clocks can have on user experience. The same will apply for multi-threaded workloads that don’t need to run continuously, like image manipulation in Lightroom or Photoshop, collages, complex Office macros, etc. Intel continues to find ways to bridge the gap between strong single threaded and strong multi-threaded performance.
I am still trying to get my hands on an 8th generation system to see how this works in practice, but in theory, it leaves little doubt that Intel will continue to dominate in this space. Intel claims we will see more than 80 systems in the market by the holiday season using these 8th generation processors.
I don’t yet have an answer from Intel as to the pricing of the 8th Generation Intel Core Processors and how they will compare to the 7th generation. If Intel plans to keep both lines going, it would make sense that Intel would raise the price of these quad-core 15-watt processors to compensate. However, if Intel is more aggressive, and wants to push the higher core count solutions to the market with its full weight, it will closely match the pricing of the 8th generation 4-core CPUs to the 7th generation 2-core CPUs.
Pricing showed up on the Intel ARK database and for now only indicates a slight upcharge for the quad-core processors over the comparable dual-core options that are shipping today. In truth, pricing for mobile processors is very different than what you might be used to for consumer hardware with bulk deals, agreements based on co-marketing, etc. so we can only tell that Intel is at least starting from a similar place with the new 8th Generation CPU pricing.
Now, let’s touch on the secondary bit of information from today’s release: the 8th Generation Intel Core Processor family will be stretching across platforms (as it has traditionally done) as well as architectures and even process technologies. Those second two bullet points are brand new to the Intel branding scheme. In previous iterations, everything that was labeled “4th generation” or what have you would be based on the same CPU microarchitecture, the same architecture design, and the same process node technology. The 6th Generation Intel Core Processor family was all based on the Skylake architecture and was built on the 14nm process node. This meant that developers and OEMs had reasonable expectations of feature support and capabilities across the spectrum of 4.5-watt processors up to the performance consumer desktop space.
For what Intel is calling the 8th generation, the company has explicitly stated that it will feature multiple architectures, including the Kaby Lake refresh processors (like the 15-watt CPUs announced here), Coffee Lake designs, and Cannonlake designs. It will also have parts built on the 14nm process technology node currently in use today as well as the 10nm process technology due next year. Coffee Lake will likely find its way to the desktop consumer market later this year as a 14nm design while Cannonlake could be the 10nm refresh of Coffee Lake and span many different product segments.
In truth, it’s all very confusing, and unnecessarily so. Intel tried to justify the change by telling us that the generations will be viewed “through the lens” of consumers and business that have specific needs for performance, battery life, and features. To me, it seems that simply excluding the Kaby Lake refresh processors from the 8th generation branding would have solved much of this issue, but maybe other Kaby Lake-based processors are still coming to fruition that we aren’t aware of that will need the branding advancement.
Regardless of what Intel calls them and why, the 15-watt processor announcements today will have a dramatic impact on performance for consumers that fall into the 2-in-1 or thin and light landscape. As a user in this space myself, I can tell you I am personally excited at the idea of better performance for light video editing on the go (even if plugged into power) while no longer needing to be bogged down by a big, heavy notebook to ensure top performance. This move should put AMD’s Raven Ridge on notice, as the expected quad-core design will now have to rely nearly completely on its integrated GPU performance to stand out. It also puts Qualcomm on edge with the pending release of Windows 10 devices powered by Snapdragon that may have more trouble dealing with potential performance considerations.
As I said at the beginning, if you thought this fall was going to be a snooze-fest after the excitement of the summer, you are sorely mistaken.
Meh, I’m not upgrading until
Meh, I’m not upgrading until I can get this performance with passive cooling.
you do not consider 15W to be
you do not consider 15W to be in the passive cooling range? Or do you mean like passively cooled within a tiny package like a thin laptop?
For desktops I consider 25W to be comfortably coolable with passive heatsinks (which is why I love the cheap amd 5350 😀 )
The recommended customer
The recommended customer price on intel’s site is $409 for the i7-8650u, $297 for the i5-8350u.
Not sure how this equates to msrp.
Don’t worry Ryan. Things
Don’t worry Ryan. Things would get pretty slow when the sunny days for vacations are over and the boring season of rains starts :p
On topic.
– So many processors announced/leaked from Intel these past days (if we also add Ice Lake) that you can NOT buy today.
– It’s obvious that Intel tries to avoid a second Ryzen on laptops.
– Intel is going rebrands full speed. No matter if Intel will have a new process node or a new architecture to show every year in the future, we will be getting a “new” generation of Core processors every year.
Will we ever see
Will we ever see ultrabook/2-in-1 with Ryzen mobile? The only laptops I can find here in Norway are two Asus RoG 17″ laptops. Will i be just another “paper launch” in the sense that no one builds the machines?
I’ve heard/read that amds
I’ve heard/read that amds mobile ryzen (Raven Ridge) will arrive before the end of the year in notebooks like 2 in 1s.
They are rumored to come in quad core, 8 thread variations for ryzen 5 and 7 (what a surprise that now, of all times, intel decides to double the core count for 15w parts now that amd is about to release the same)
Note, these raven ridge parts are NOT like that full on desktop ryzen chip used in the 17 inch gaming notebooks, these are apus, and like intels 15w mobile chips, will come with integrated graphics.
Except these are reported to come with Vega integrated graphics, so it will be interesting to see how that stacks up with intels integrated graphics. At lower power levels, Vega should be more efficient as it’s not trying to chase the dragon of nvidia. If I had to had to guess, I suspect the baseline graphics performance of amds raven ridge chips will be higher than intels 620 integrated graphics, but perhaps a bit lower than intels iris pro 650 graphics (though those pull more power and are NOT on the 15w parts). I’d be a bit surprised if amds integrated graphics is not higher performing than intels at 15w, but we’ll see.
AMDs raven ridge does not have to beat intels 8th gen, they just have to get close at a lower cost with comparable battery life. Get in some of the more more premium notebook thin and light designs, for 100-200 dollars less money and I think PLENTY of people would be happy to go amd instead of intel for their notebooks.
Look at that Base Clock 1.9 –
Look at that Base Clock 1.9 – Max Turbo Clock 4.2 GHz on that Core i7-8650U, that’s quite some range but what are the all core rates. I’d like to see that Blender 3D benchmarks run on this SKU. I Hope this 4 core/8 thread offering at 15W becomes a 6 core/12 thread thread SKU at 35W+ on some regular from factor laptop variants.
And maybe the very next year some 4 core/8 thread i7 SKUs will once again be on clearance at Microcenter at a good price. Bucause currently there are not many laptops, HP probooks, etc. that offer more than dual core i7 U series SKUs. I’ll be looking for some Raven Ridge APU’s to show up in laptops come the end of the year and hopefully there will be 35W+ laptops variants offered also for the power user. Fatter laptops with better cooling solutions are as rare as it gets what with all the thin and light obsession folks.
If you want thin/fanless then by all means get a tablet and a keyboard combo but really there needs to be some more powerful laptop offerings without having to go into gaming laptop territory. I liked the Probooks because they had usually came with AMD discrete mobile GPUs/core i7 QM offerings at a more affordable price point but not now for many i7 “QM” variants. And those dual core i7’s are a deal killer and quad core i7’s SKUs of the Ivy Bridge-Broadwell generation have all but disappeard from laptop offerings since Skylake’s arrival on the retailer’s shelves. It’s mostly been dual core for i7 on down on laptops over the last few years and it’s hard to find many laptops with the more affordable AMD APUs(with dual channel DDR4 memory) or discrete AMD graphics in a laptop(35 watt) form factor design.
Given that most reviews of
Given that most reviews of Kaby lake laptops suggested that the i5 and i7 processors ended up with identical performance due to thermal throttling one wonders how effective quad core parts will even be.
That’s on the Laptop OEMs and
That’s on the Laptop OEMs and their design skimping wth regards to thermal performance on their laptop SKUs. And given Intel’s market control with Intel’s pricing eating into all the Laptop OEM’s already thin margins, I’ll expect that practice will contine with Laptop OEMs doing more poorly designed laptops that come with more bloatware because Intel is taking all the profits and the Laptop OEM’s can do nothing currently about that.
Now maybe there can be some Raven Ridge competition but the Laptop OEMs will be under great pressure if Raven Ridge, as demonstrated in some AMD refrence laptop builds, does not show outstanding performance relative to Intel’s offerings. The Zen CPU Micro-Arch is great but that Vega integrated Graphics is going to have to beat all of Intel’s graphics offerings for at least a high end Raven Ridge mobile SKU that Intel can not touch in graphics performance.
AMD needs to double down and get some Raven Ridge mobile Laptop APU offerings that come with HBM2(Single stack of 2GB or 4GB) so Vega’s HBCC/HBC IP can be utilized. AMD’s current 8GB Vega discrete offerings have so much VRAM that most games can not be used to stress test that Vega HBCC/HBC IP currently! But that HBCC/HBC IP will really shine on mobile Vega GPUs with smaller 2GB/4GB VRAM sizes and current games will be able to stress the Vega HBCC/HBC IP and force the HBCC to make use of the laptop’s regular DIMM based DRAM for Virtual VRAM under the HBCC’s control.
Go over to techgage as there is a new article where the Author is currently trying to do some benchmarking in order to stress Vega 64’s/56’s HBCC using games that are so optimized for 8GB VRAM that the VEGA SKUs tested can not be made to make use of the HBM2-Cache mode enough for proper testing to be done under that Vega HBCC/HBC IP that comes with Vega. And there is no BIOS switch to disable HBM2 down to less than 8GB currenty.
The upcoming Raven Ridge APU SKUs, if they have no HBC/HBM2 for the HBCC to make use of as HBC, are going to have to rely on the Laptop OEMs all providing more than a single channel to DDR4 DRAM or the Vega nCU’s/shader cores will be starved for necessary bandwidth once again on laptops!
I’m fairly certain that amds
I’m fairly certain that amds raven ridge will not be gimped to single channel memory configurations like Carrizo and everything else on the notebook front. That will give more memory bandwidth from ddr4 to feed the integrated gpu. I don’t know if amd is going to be able to take out intels higher end iris pro on notebooks that use 28+ watts of power, but like I said above, I think amd has a really good shot at blowing past the integrated gpu performance of what MOST intel u series parts come with, the 620 graphics.
The thing that I find most
The thing that I find most engaging is the idea of 4 core / 8 thread laptop processor in a thin and light combined with external graphics. If I recall correctly when the external graphics enclosures were tested, the limiting factor was CPU.
Ryan, what are your thoughts on this? Will these new 4/8 15 watt CPU’s make the external graphics card enclosure a viable alternative to a gaming desktop? Would 4K gaming be possible on them?
What was the version of
What was the version of Thunderbolt used for the benchmarking? And also IF TB3 still can not deliver enough bandwidth to a GPU(Middle/High Range GPU SKU) then how can that the fault of only the CPU. There still has not been enough testing with TB3 and certianly a 15W CPU SKU is not going to be able to 4K game as that requires a high end desktop CPU. And even if the Laptop had a desktop processor inside that TB3 40Gb interface is going to be taxed to its limits and choke off some GPU’s performance.
over at ultrabookreview website they list laptops with TB3 but they caution:
“There’s a catch about the Thunderbolt 3 connection, it can be hooked up to the motherboard via x2 or x4 PCIe lanes. The x2 TB3 port are half the speed of the x4 TB3 connection and is primarily a problem when hooking up an external graphics chip and playing on the notebook’s internal display.”
So TB3 and 4K(With all the settings on Low/Mid) on an external monitor, maybe, but not when trying to port the graphics back from the external GPU to the laptop’s display over the same TB3 bandwidth. It’s too bad that TB3 does not come with a link aggregation ability so 2 TB3(Full 40Gb effective 80Gb) cables cuuld be run to an external GPU box. But a laptop SKU with the 8 extra full PCIe 3.0 lanes available that it would need to used to supply the necessary PCIe 3.0 connectivity to feed 2 TB3 full 40Gb ports would not easy to find, and laptop motherboard SKUs are not known to have that much extra PCIe/Lanes resources to begin with.
PCIe 4.0 can not get here soon enough as far as laptops are concerned and TB4 better be incoming also.
What would be nice for laptops to have also is a HDMI, or Display Port IN connection for laptops. That way the CPU to GPU(on an external thunderbolt box) bandwidth would not be taxed at all by running the external GPUs output out the GPU’s DP/HDMI and back to the laptops display via HDMI/DP bandwidth over a native HDMI/DP one way IN-Connection on the laptop.
I know these are configurable
I know these are configurable TDP, but I wish more OEMs made 20 watt parts. I’d gladly take an extra bit of thickness for the major clockspeed boost that would provide.
Sorry Ryan, this one had me
Sorry Ryan, this one had me doing mind flips around a double negative.
“This not a new architecture nor is this is not a new process node”
“This is not a new architecture nor is this is a new process node”
Unless it is on a new process node 😀 Thanks for the great roundup on this 7th / 7 1/2th gen information release.
You mean 6 3/4
You mean 6 3/4 generation?
This “8th-gen” branding is getting as bad as the GPU industry. “7th-gen” Kaby Lake had absolutely no architectural improvement in the CPU; it was a tweaked 6th-gen CPU and an integrated GPU with an added decoding feature. Now we’re looking at an “8th-gen” CPU that is specifically listed as Kaby Lake.
Intel can try to explain this however they want. The reality is they need to give the impression of market leadership. Sad, Intel. Just pathetic.
UHD graphics, so now you
UHD graphics, so now you simply must benchmark them at UltraHD resolutions, and reveal their 2-5 fps goodness.
This is totally a snoozefest.
This is totally a snoozefest. Its a single new die on a single new package and a big marketing flurry.
I detest these BGA boards.
I detest these BGA boards.
Beware that Intel will be
Beware that Intel will be releasing Crystal Lake processors on Friday the 13th
Announcement from Jason
Announcement from Jason Voorhees the new head of marketing at Intel. Crystal Lake processors will be released Friday the 13th