Intel’s 14nm Core M Series “Broadwell-Y” Processors With 4.5W TDP Detailed – Powers Fanless and 2 in 1 Mobility Devices
Intel has dedicated the last few months to the desktop market by launching two new platforms for the PC desktop community. In April, we saw the Haswell Refresh platform which was launched along side a flood of 9-Series chipset based motherboards. In July, Intel introduced their latest Devil’s Canyon processors and just last month, we saw the daddy Haswell platform codenamed Haswell-E which ushered a new era of enthusiasts level PCs with the latest Haswell-E processors, X99 motherboards and the first platform to support DDR4 memory on the consumer front.
Intel’s 14nm Core M Series “Broadwell-Y” Processors With 4.5W TDP Detailed
With the majority of Q2 and Q3 targeted towards the desktop market, Intel is now making a return on the ever growing mobility front. Intel is not going to release a new Haswell processor since they are done with that CPU architecture and Haswell-E is the last major platform release on the 22nm Haswell micro-architecture. Intel is now expanding to the mobility front with their latest 14nm process based Broadwell architecture with the first cores configured specifically for the mobility market which include 2 in 1s, notebooks and fanless devices. Meet the Intel Core M processor.
The Intel Core M name isn’t new and we got a glimpse of the processor a few weeks back when Intel gave the green light to publish some of the key architectural details regarding their new 14nm processor node. It’s very essential that we take a look at the architecture first before we get to talk about the Core M processor in detail. Intel’s focus with Haswell was to produce a highly scalable architecture which spans from tablets to desktops and even the HPC workstation market. They are keeping the same focus with Broadwell which will power a lot of devices and form factors. While PC is an established market, Intel will puts its core focus towards the growth markets which include notebooks, tablets and mobile smartphones. Intel has already a variety of Ultrabooks in the market and they recently entered the 2 in 1 form factor design which has got some attention but there are some specific rules and regulations which need to be followed. As a part of the standards which define an Ultra book, Intel wants to keep these devices power efficient, performance capable and feature days of battery life without getting hot.
Haswell has made impressive achievements in every field related to efficiency. Although it failed to deliver on the high performance promises on the desktop side, but the chip actually proved well for the mobility side of things. Being on the 22nm Tri-gate transistor technology, Haswell introduced 2 times the battery life, 20 times the idle power reduction, MCP (Multi-Chip Processor) integration, FIVR (Fully Integrated Voltage Regulator) and an increased dynamic operating range. The next evolution in the low power CPU roadmap is Broadwell-Y also known as Core M. Broadwell-Y replaces Haswell-Y and Ivy Bridge-Y series before it. Some of the key features of Broadwell-Y (Core M) are the use of 14nm Tri-Gate transistors (second generation), reduction in TDP, thermal and power management system integration, reduced SOC idle powers, increased clock operating range as per thermal/power requirements, use of 2nd generation FIVR with the help of 3D Layering technology, new GPU and CPU core for better performance and a SOC chipset design which focuses on lower power which enhancing the feature set of the Core M series processors.
The 14nm Process Node – 2nd Gen Tri-Gate (FinFET) Transistor Technology
Yes, Intel’s Core M is the first consumer level processor to make use of the latest 14nm process node. There has been much rumor and reports regarding the delays of Intel’s 14nm which have been true since this launch was actually going to take place earlier this year however moved to Holiday 2014 due to production issues in Intel’s fabs. However, after all this time, we finally have a first look at the 14nm Broadwell processor which has now been qualified for volume production. The 14nm processor node makes use of the second generation tri-gate (FinFET) transistor technology which deliver better power, performance, density and cost per transistor. Just like I mentioned previously, Broadwell will be used to manufacture a range of products from high performance to low power. The design of each transistor fin is now more denser and compared to the 22nm node, 14nm is much smaller and efficient.
Looking at the transistor fins on the 22nm node, the height of each transistor fin is 34nm with a silicon substrate base and a 60nm pitch in the fin layout. Intel has managed to reduce the pitch distance to 42nm on their 14nm process design which makes it more tighter hence improving the density. But the pitch hasn’t only gotten tighter, but it has also been made taller and thinner. So each fin now measure 42nm in height compared to 34nm on 22nm node. This results in increased current flow and performance and the height and pitch have now been set to 42nm each while lower fins mean further improved density and lower capacitance. The fins are topped by a second generation Metal Gate and a close up shot of the fins can be seen in the gallery below. The interconnect scaling has also been updated with a new design with the interconnect path now boasting a 52nm minimum pitch which provides better scaling compared to the 80nm pitch on the 22nm process node.
These architectural improvements enhance the per transistor performance and deliver lower leakage so that the product is more performance and efficiency focused then previous generations. The lower leakage power also means that Intel can now deliver more performance on a wide range of products, especially the mobility devices. A better graph posted shows a 5-6x performance per watt on server, laptop and mobile applications with the 14nm design while active power is reduced to .25x on mobile and performance on the server side is impacted by a 2 x increase compared to 45nm technology. Each passing generation since 45nm, Intel gained a 1.6x performance per watt improvement. Compared to Haswell, that has been shifted to a 2x performance per watt gain with the new Tri-Gate transistor technology, better area scaling on the 14nm node, enhanced process optimization. Intel also mentions that their logic area scaling has continued to scale 0.53x per generation and while other’s have tended to have better density in the past, they came later and now have switch their focus to FinFETs from Planar.
The current Broadwell-Y SOC will set the standard for future Broadwell releases but its not the end in terms of innovating. Intel promises to add more improvements in future releases and will be introduced in chips under going development. Intel’s CEO, Brian Krzanich, has also confirmed that Intel will not use EUV lithography for their future 10nm process node which is due to arrive on 2016 and that they can get through 10nm without EUV. Next up, we detail the architectural bits of Intel’s Broadwell microprocessor codenamed Core M.
Intel Core M Processor Microarchitecture Details:
Efficiency, Low Power, 2nd Gen FVIR and 3D Layering Technology
Before we take anything into consideration, the Broadwell microarchitecture we are currently looking at is part of the Core M processor which is a low power optimized core. While Core M is the reference design for future Broadwell SKUs, it should not be compared to the high performance releases which we will get later on in the Core series line for high-performance notebooks, desktops, servers and workstations.
The Intel Core M “Broadwell-Y” has been optimized for mobility devices but more over Intel has set a few challenges ahead for this launch. They want to make processor which can be used on fanless devices which they term “The Fanless Challenge”. Did the challenge pass? It definitely did and we are already looking at a spree of devices at IFA 2014 from Lenovo, Dell and ASUS which feature the Core M processor and have the fanless design. This is good news but let’s detail the micro architecture first before moving into device details.
So some of the key aspects of the Broadwell microarchitecture would be to be suitable for fanless devices, deliver on core performance improvements and feature enhanced graphics, media, display capabilities. There are different devices with various sizes and various power standards. The 10.1 inch market which scale from thinness of 8-10mm are suitable for fanless operations and power consumption of 3-5W. The larger size means that you could integrate some higher power demanding chips with better cooling but the low form factor and mobility designs require a fanless approach. The same design can be used for notebooks if it delivers on the promise of better performance. The 14nm process technology is just a part of this promise that needs to be integrated into a processor capable of running these devices.
Broadwell-Y has several enhancements on this side, the SOC delivers lower capacitance that enables 25% lower power input, 10% lower minimum operating voltage that saves an additional 20% power, 10-15% performance per transistor increase through the use of low-voltage optimized 14nm transistor technology, 2 times lower leakage which means it requires 10% lower power since most of the power is utilized instead of being leaked and better area scaling.
An SOC comparison shown by Intel reveals the different in terms of packaging and size of Intel’s Haswell-Y and Broadwell-Y series SOCs. Compared to Haswell, Broadwell-Y is 50% smaller in terms of height and width while being 30% smaller in breadth. The exact dimensions of both SOCs are 40x24x1.5mm for Haswell-Y and 30×16.5x 1.04mm for Broadwell-Y. The entire PCB area reduction is around 25% compared to Haswell-Y. Intel is so much focused on the power output of the Core M die that they even designed a new PCH known as Broadwell PCH-LP which is focused on power. Located on the same die, the Broadwell PCH (Low Power) comes with a audio DSP featuring advanced post processing for audio and PCI-e storage options. This tones down the idle PCH power down to 25% compared to last year Haswell-Y parts and reduces active power by 20% compared to Haswell PCH-LP.
Another key enabler of the Core M processor is the use of 2nd generation FIVR and 3DL for power efficiency delivery. We all know that Intel is going to phase out FIVR when Skylake hits the market but Intel is keeping it alive for one more generation and with 2nd generation fully integrated voltage regulator that is used on Broadwell, the Core M processors deliver better efficiency through non-linear Droop control and dual FIVR LVR modes. Intel has also used a new technique to relocate inductors from the main die to integrate modules under the die using 3D Layering (3DL PCB) which save power. All these technologies add up to deliver lower power output on Broadwell generation of processors including the new power gating and boost algorithms which maximize frequency output while conserving clock speeds for power when under idle mode. To sum this all up, Intel’s Core M CPU which is Broadwell-Y SKU is going to be power efficient and deliver several features and enhancements in the mobility market.
Update: Computerbase had the courtesy to show a few comparisons of the Core M based motherboards. The comparison show a Core M device board against a Macbook Pro board which features Haswell-Y chip. There’s a major difference in size between both chips and we can see how this aim will improve products designs for tablets.
There’s a new turbo boost algorithm in place too which maximizes the boost clock opportunity for the processor. There are three key points, PL1, PL2 and PL3. The maximum boost the chip can hit is during PL3 and the time depends on the set limit which has to protect the battery from overloading during this limit but when there is enough power at hand, the CPU will prioritize the cores for maximum performance. The PL2 is the standard burst limit while PL1 is the long term system limit which is defined by standard. Intel has also set a new system optimized thermal management and regulation unit which takes all key components of the device that manages and balance the power and thermal framework of the Core M based devices.
Intel Broadwell Converged Core – The CPU, The GPU and The PCH
Intel’s Broadwell is basically a Tick which means its a die shrink of a new architecture that was previously intro’d (Haswell = Tock). Based on a smaller, 14nm node, Intel’s Broadwell has managed to integrate new features aside from just core enhancements. Given Intel’s slow generation over generation IPC improvement, we are looking at only 5% increase in instruction per clock compared to Haswell which falls in line with Ivy Bridge and Haswell. Intel hasn’t gained significant gains in the IPC department since a while but this model works for them.
The Intel Core M die shows a block diagram of the flagship Broadwell-Y series processor. The chip features a total of 1.3 Billion transistors while Haswell-Y series featured a total of .96 Billion transistors. The Die size is 82mm2 compared to 131mm2 on Haswell-Y which shows the PCB space management I was talking about on earlier. The majority of the die is dedicated to the graphics core while there are two CPU cores and right below is a shared 4 MB L3 cache which is shared across the graphics and processor cores. There’s a small section dedicated to the memory and I/O controllers at the bottom of the die while along the right side, we can note the system agent, display engine and memory controller.
Tweaks that help deliver these enhancements include larger schedulers and buffers, larger L2 cache with dedicated 1 GB page L2, Faster floating point multiplier, Radix-1024 divider, improved address prediction for branches, faster round-trips for visualization and targeting cryptography acceleration instruction improvements. These enhancements will set the basis of the 5% IPC gain over Haswell while power efficiency will see new power gating and design optimizations. A 5% gain over Haswell is decent and since these core enhancements will scale across the entire Broadwell CPU family, we can expect the same from the Broadwell desktop processors which arrive in Q1 next year.
Over in the GPU department, Broadwell will feature several architectural enhancements with increase execution units of 20% which means we are looking at a EU count of 24 on Broadwell HD Graphics 5300 chip compared to 20 on Haswell-Y series. The Core M series chip will be clocked up to 850 MHz and drive resolution of up to 3840 x 2160 (4K). It will also feature increased sampler throughput of 50%. So for desktop parts this means that iGPU enabled CPUs which featured 40 EUs will be upgraded to 48 EUs on the Broadwell microarchitecture delivering faster clock speeds thanks to more thermal headroom provided on the 14nm process, increased geometry, Z, Pixel fill rates and enhanced performance per core.
Software side improvements will include continued focus on gaming with support for DirectX 11.2, DirectX 12, OpenGL 4.3 and Open CL 1.2 / 2.0 for GPU compute. Intel’s own Quick Sync Video, Clear Video HD and wireless display technologies will be available through the HD 5300 graphics core. The GPU side will also feature a hybrid styled H.265 decoder which will use shaders of the iGPU to decode and execute the decoding process. While this process is useful over the traditional CPU, it is less power efficient. The Broadwell Y-Series CPU will feature 4K video support, HDMI 1.4 and Display Port 1.2 / eDP 1.3a.
The Intel Core M CPU Lineup – 5Y70, 5Y10a, 5Y10
Intel has three Core M “Broadwell-Y” series processors in the stack, the 5Y70, 5Y10a and 5Y10. The Intel 5Y70 is the flagship processor which features 2 cores and four threads. The processor maintains a core clock of 1.10 GHz offering a boost of 2.60 GHz in single and dual core modes given the power and thermal limits. The chip boasts the Intel HD Graphics 5300 chip which includes 24 execution units clocked at a base clock of 100 MHz and maximum frequency of 850 MHz. The chip supports 1600 MHz LPDDR3L and LDDR3L memory. It comes with 4 MB of L3 cache, a 4.5W TDP and a thermal limit of 95 degrees (Celsius).
The Intel 5Y10a features 2 cores and four threads. The processor maintains a core clock of 800 MHz offering a boost of 2.00 GHz in single and dual core modes given the power and thermal limits. The chip boasts the Intel HD Graphics 5300 chip which includes 24 execution units clocked at a base clock of 100 MHz and maximum frequency of 800 MHz. The chip supports 1600 MHz LPDDR3L and LDDR3L memory. It comes with 4 MB of L3 cache, a 4.5W TDP and a thermal limit of 95 degrees (Celsius).
The Intel 5Y10 features 2 cores and four threads. The processor maintains a core clock of 800 MHz offering a boost of 2.00 GHz in single and dual core modes given the power and thermal limits. The chip boasts the Intel HD Graphics 5300 chip which includes 24 execution units clocked at a base clock of 100 MHz and maximum frequency of 800 MHz. The chip supports 1600 MHz LPDDR3L and DDR3L memory. It comes with 4 MB of L3 cache, a 4.5W TDP and a thermal limit of 95 degrees (Celsius). The Intel 5Y10 is the only chip in the stack with a configurable TDP design that can be toned down to 4W.
Intel Core M (Broadwell-Y) Series:
|Core M 5Y10||2 / 2||800 MHz / 2 GHz||4 MB||100 / 800 MHz||DDR3L/LPDDR3-1600||4.5W|
|Core M 5Y10a||2 / 2||800 MHz / 2 GHz||4 MB||100 / 800 MHz||DDR3L/LPDDR3-1600||4.5W|
|Core M 5Y70||2 / 4||1.1 / 2.6 GHz||4 MB||100 / 850 MHz||DDR3L/LPDDR3-1600||4.5W|
Intel Core M Series Discontinued Prior To Launch?
New developments also confirm that Intel has officially discontinued their first Core M series lineup prior to its launch. In a presentation, Intel has stated that the three Core M processors, namely the 5Y70, 5Y10 and 5Y10a have been discontinued and while products will ship with the current chips, a new revision which is rumored to be the F0/E0 will ship later on in the devices. The devices based on the affected Core M chips will either be replaced or offered with new firmware updates close to launch.
Why is there a need for Core M – Meeting Users Demands
We all know that desktop is an established market and is not in much momentum (on the OEM front) as the mobility market. No doubt that the PC desktop market is constantly gaining buzz in the DIY field but the OEM makers have to look for new devices to serve their consumer demands. People these days have switched to mobility for their compact form factor, on-the-go usage and user friendly interfaces. All 2 in 1, tablets and notebook users are constantly upgrading their devices and Intel mentions that 2 in 1 buyers refresh their PCs 1 year earlier than notebook buyers since they want updated devices with more features. Intel has planned a 3x increase in the amount of 2 in 1 systems with up to 70 designs planned for launch in Q3 2014. Similarly, the cost of these devices will be <$699 US.
Intel Core M seems like the perfect solution for these devices which need to be fanless, slim, weightless and easy to carry. Since Core M is the first 14nm device, it will meet the demands of next iteration of mobility devices boasting better performance and efficiency. The Core M processors will stick to the fanless range of 4.5W and even down to just 4W with configurable TDPs. The devices include Detachable 2 in 1, Convertible 2 in 1, Notebooks with touch, Traditional notebooks, Enthusiast notebooks, Mobile workstation and premium tablets. Some enhancements over a 4-Year old notebook include 11.6″ touch screens, resolution of 1920 x 1080 (Finally!!!), 3 times thinner tablets, 50% lighter designs at just 1.76 lbs, SSD and HDD storage, Faster LAN (802. 11 AC WLAN), and USB 3.0 connectivity for 10 times faster transfer speeds.
The main comparison however is between the 2 in 1 form factor itself against the tablet which Intel says is the ultimate rival of their new 2 in 1 form factor. Starting off, Intel has some comparisons to differentiate their PC form factor such as it boasting the same portable thing design with extended 8 hours plus battery life and less than 2 lbs weight. Both feature basic content creation, users can watch videos and enjoy music. But here’s the differences, the 2 in 1 can actually play and run both tablet and PC games/applications while tablet is limited tot its own applications. Similarly, there’s greater room for Multi tasking with several windows, multiple file format support and excess of storage to hold. A premium tablet which is compared can be grabbed fro around $900 – $980 and that’s without adding the prices for additional accessories. The price of the 2 in 1 devices have yet to be determined but the recent figures do pit the prices close to the $900 – $1000 US mark that is similar to the premium tablet offerings and plus for better processors, memory, storage, etc.
Intel has some performance numbers they want to share which pit a 4.5W Core M “Broadwell-Y” based 5Y70 (2.60 GHz, 2 Cores / 4 Threads, 4 MB Cache) processor against a Arrandale based Core i5-520UM processor ( 1.06 GHz, 2 Cores / 4 Threads, 3 MB Cache) which featured a TDP of 18W. Intel basically wants to show the performance difference between a 4 Year old equivalent PC with performance increases of 2.1x in office productivity benchmarks, 7.1x in 3D graphics, 2.3x in web applications, 7.6x in video conversion, 1.7x in word to PDF conversion, 1.9x in photo editing and all of that at 25% of the power of the Core i5-520UM processor.
Last up, we have a few power and battery life numbers directly from Intel based on the 35Wh battery configuration. There are several improvements in SOC, Panel and the entire platform numbers compared to Haswell-Y but the most biggest difference notable is the audio section which demands little to no power input at all. There’s a good 73.5 minute average battery life gain over Haswell-Y which is impressive. So there you have it, these are all the Intel Core M details you need to know and now we should focus on the newly introduced Core M devices.
Intel Core M – The Devices
At IFA 2014, several manufacturers have revealed their new Intel Core M based 2 in 1 devices featuring the fanless designs. Some of the notable devices include ASUS’s ZenBook UX305, Dell’s Latitude 13 7000, HP Envy X2 13/15 and Lenovo’s Thinkpad Helix 2 in 1. More detailed info is provided below:
The ASUS Zenbook UX305 is the latest product in ASUS’s Zenbook lineup featuring the Core M processor. The Zenbook UX305 boasts a 13.3″ display featuring a resolution of 3200×1800 (QHD+). Since it is based on the M Core processor, the Zenbook UX305 features three USB 3.0 ports, an SD card reader, Micro HDMI and 802.11 AC WiFI LAN Wireless connectivity. The notebook weighs in at just 1.2kg and measures 12.3mm in thickness (0.48 inches). The Zenbook will be available in two colors which include Ceramic Alloy and Obsidian Stone Its price has not been announced yet but ASUS did confirm that the model will ship with 128 / 256 GB of SSD based storage, 8 GB of DDR3 system memory along with the Core M processor which may include the Core 5Y70.
Moving to HP, the company plans to launch two HP Envy devices based on the Core M series processors. The 2 in 1 fanless PCs will feature a 13.3 inch screen (HP Envy X2 13) and 15.6 inch screen (HP Envy X2 15). Both note books feature touch screen display and boast a resolution of 1920x1080P (Full HD). Both devices feature a Core M series processor with a 0.9 GHz base clock and 1.6 GB turbo boost frequency. Neither of the current lineup boasts such frequencies so its possible that the Core M F revision which is now under production may become part of the HP ENVY X2 series. The original Core M series lineup has been discontinued but the chips have been shipped to OEMs and will be featured in the current devices with the latter revision based products arriving a bit later in November 2014 time frame. Back to the specifications, the HP ENVY X2 13/15 both feature 4 GB of DDR3 memory along with the choice of either 128 GB of SSD storage or 500 GB of SHDD. The pricing suggested for the Core M based devices is up to $699 US however the HP ENVY X2 were listed for € 899 which is quite alot for these products.
The Dell Latitude 13 7000 is another 2 in 1 form factor device based on the Core M processor. The Dell Latitude 13 7000 is based on the first 14nm consumer processor and boasts a 13.3 inch display capable of full HD res of 1920x1080P. The 2 in 1 device features a unique set of wireless docking capabilities which will be detailed later on. In terms of specifications, the laptop will feature up to 8 GB of DDR3 memory which means price optimized variants will also be available in 4 GB DDR3 models. Storage capacity up to 512 GB will be available however its not mentioned whether it will be based on standard HDD or SSD solution. Dell includes optional LTE broadband and wireless capabilities thanks to 802.11 AC WiFi. A front 2 megapixel and 8 megapixel rear facing camera can be located on the main detachable tablet. Pricing is suggested around $1199 US.
The final device in detail is the Lenovo Core M based 2 in 1 Thinkpad Helix notebook. The Thinkpad Helix 2 is a premium priced windows tablet and notebook (2 in 1) which comes in at a hefty starting price tag of $999 US. The model comes with 802.11 AC WiFi wireless connectivity and weighs in at just 1.76 pounds while being only 0.38 inches thick. In terms of specifications, the Core M based device is completely fanless and comes with a keyboard dock which pushes the weight to 3 pounds. The device comes with up to 8 GB of DDR3 memory and choice of either HDD or SSD storage.