Intel is planning to update its rather successful NUC (Next Unit of Computing) series and as you can expect, they will come with Broadwell CPUs inside.
Intel isn’t hiding the external design of the new cases and there is a dominant yellow connector at the front of the new NUC, and this one should be providing charging power even when the device is turned off.
The board comes with either M2 storage or single SATA and there will be two different designs one exclusively for M2 drive and the second taller that will be able to take 2.5 inch SSD or HDD as well.
We will probably learn more details at CES 2015 that is about to start in less than three weeks from now, but the Broadwell in this small form factor will get a speed boost and some future prove technologies such as M2 SSD support.
We are running Core i5 4200 powered NUC with Windows 10 and it really works great powered by 240GB Kingston mS200 mSata SSD and Impact SO DIMM memory. These machines takes less than half an hour to assemble and boot into windows, including Windows 10 and make a perfect choice for the lovers of quiet computing.
The new version will obviously run at least slightly faster than the one we are testing and the marketing is excluding about “the one with the yellow USB connector”.
A company insider has spilled the beans in Korea, claiming that Samsung has started Apple A9 production in 14nm FinFET.
The A9 is the next generation SoC for Apple iPhone and iPad products and it is manufactured on the Samsung – GlobalFoundries 14nm FinFET manufacturing process. In the other news, Samsung’s Ki-nam, president of the company’s semiconductor business and head of System LSI business has confirmed that the company started production of 14-nanometre FinFET chips.
The report mentions Austin as a possible site for Apple products but we wonder if the GlobalFoundries Fab 8 in New York State could become one of the partners for the 14nm FinFET manufacturing. Samsung didn’t officially reveal the client for the 14nm FinFET, but Apple is the most obvious candidate, while we expect to see 14 / 16nm FinFET graphics chips from AMD and Nvidia but most likely in the latter half of 2015 at best.
Qualcomm is likely to announce new LTE modem based on 14nm FinFET and the flagship SoC Snapdragon 810 is a 20nm chip. Qualcomm is manufacturing its 810 chips as we speak to meet demand for flagship Android phones coming in Q1 2015. Flagship Samsung, HTC and LG phones among others are likely to use Snapdragon 810 as a replacement for this year’s Snapdragon 801, a high end chip that ended up in millions of high-end phones.
Samsung / GlobalFoundries14nm FinFET process is 15 percent smaller, 20 percent faster, and 35 percent more power efficient compared to 20nm processors. This definitely sounds exiting and will bring more performance into phones, tablets, GPUs and will significantly decrease power consumption. The move from 28nm is long overdue.
We believe that Qualcomm’s LTE modem might be the first chip to officially come with this manufacturing process and Apple will probably take most of the 14nm production for an update in its tablets and phones scheduled for 2015.
Samsung is having another crack at building a GPU.
This is not company’s first attempt to make a GPU and this time it is meant to be used with its SoC and not in graphics cards. Samsung has announced last year that it wants to make its System on Chips based on in-house 64-bit architecture but we still have to wait and see one eventuate.
Samsung is trying to make a GPU for years and enter this already crowded GPU IP market. Qualcomm uses Adreno, Nvidia uses Geforce and wants to license it to others. Apple uses PowerVR while Mediatek uses ARM owed Mali graphics for newer processors while using PowerVR for some older parts. Intel is using PowerVR G6430 for its mobile processors such as Atom Z3580 Moorefield while AMD has its own graphics that it can use for future SoCs and APUs. Intel owns Intel HD graphics that dominates the integrated CPU market especially for notebooks.
Samsung currently uses Mali graphics but this might change. If its team is successful, it might come with its own graphics and jack them under the bonnet of its own Exynos processor by the next summer.
All the sudden Nvidia’s lawsuit against Samsung makes more sense.
Samsung is trying to get into Nvidia space and the company doesn’t like it. Even if Samsung manages to make a successful GPU, the competition is hard. Even with years of trying Samsung is mostly using Exynos for its own tablets and some phones. Most Samsung high end phones use Qualcomm Snapdragons as these tend to have better LTE modems and are widely available.
According to the Korean ZDnet the company might talk about the GPU as early as February at the Solid Circuits Society (ISSCC) conference with the official announcement scheduled for summer 2015.
Samsung has filed a complaint with the U.S. International Trade Commission as part of patent war which appears to have broken out between the two chipmakers. Samsung claims Nvidia infringed several of its chip-related patents and for making false claims about its products. This is effectively counter-suing after Nvidia filed a suit against the company in September making more or less the same charges.
Nvidia accused Samsung and rival Qualcomm of infringing patents on its graphics-processing unit (GPU). Samsung, which had filed the lawsuit in a US federal court on November 4, is seeking damages for deliberate infringement of several technical patents, including a few that govern the way semiconductors buffer and use data.
The ITC complaint also named computer-parts manufacturers Biostar Microtech and Elitegroup. These things run and run and usually wind up with a settlement where both sides agree to keep the details quiet. The ITC is often used as leverage in such cases because it deals with things a little quicker and a product embargo to the US can be seriously damage a company’s wealth.
Bay Trail was quite a big deal when it started shipping in late 2013.
It was a tablet chip that enabled great design wins such as the affordable Asus T100TA and even in late 2014 Asus used the platform to create the EeeBook X205, a $199 netbook.
Both of these designs are based on Intel’s Bay-Trail M processor, a year old 22nm quad-core processor based on the Silvermont design. Some machines that are coming with LTE, both netbooks and tablets and there will be new chip coming in 2015. It is called LTE Advanced XMM7360 chip and supports LTE Cat 10,3 CA up to 450 Mbits download and upload.
Intel will also offer Morrefield quad cores for machines with lower TDP ratings, especially tablets, and at some point in 2015 it will introduce its 14nm Airmont core based Cherry Trail processor. Cherry Trail based on 14nm Airmont core was originally expected in late 2014, but it got pushed towards middle of 2015.
Intel is clearly encountering more obstacles moving from the 22nm to the 14nm manufacturing process, but considering that most ARM competitors still have to start commercially shipping its 20nm SoCs in significant volumes, Intel still has a manufacturing node advantage. If only Intel had as many design wins to go along with its cutting edge fabs, as the company has been struggling to ship 40milion tablets in 2014, as promised.
Braxton will replace Cherry Trail in 2016. Braxton is a tock architecture, another 14nm design based on the quad-core Goldmont core. When it comes to the Performance Media Internet Device (MID) market Intel has another chip planned in 2016. It calls it SoFIA MID and the chip comes in intels 14nm manufacturing process.
Value and Entry markets for Media Internet Device (MID) and phones includes four new SoFIA parts, but with all these new and exciting chips Intel has to compete against some advanced chips coming on line in 2015, including the Qualcomm Snapdragon 810 20nm, Nvidia Erista and more affordable Mediatek solutions such as the MT6795 A53-based octa-core and its successor.
Intel has revealed an update to its CPU roadmap and some things have changed in 2015 and beyond. Let’s start with the Broadwell 14nm architecture. The first processors based on this technology are starting to show up in razor thin laptops and ultra-fast tablets. Intel should add ultra-expensive, as the first SKUs cost an arm and leg.
However, the Core M is excellent performer considering its 4.5W TDP envelope and this is the first time that Intel has made such as energy efficient Core processor. There will be some machines based on Core M vPro for business users and they should be available right now, depending on the market.
It almost feels like Intel launched Broadwell 14nm products just to please investors, as we don’t see too many people spending $999 or more on an ultra-thin notebook. It is reminiscent of the Ultrabook push, at least for now. Intel claims that there are 70 hybrids and two-in-one designs on the market right now. We will have to look into this number as it looks quite optimistic to us. We have to trust what Intel’s SVP Kirk Skaugen, the leader of the chipmaker’s PC group, is telling investors.
Intel has mentioned the 5th generation Core architecture and future notebooks and 2-in-1 PCs slated to launch next spring. This is a rather wide definition as spring starts in late March and ends in late June, so an educated guess would be mid- to late-Q2 2015, with rollouts at Computex 2015. Intel continues the tradition of launching Core i3, Core i5 and Core i7 processors for customers and two processors for business users, namely Core i5 vPro and Core i7 vPro products. We suspect that Intel will call them Core I 5000 generation.
It looks like Braswell, the next generation Atom core, has been delayed. Intel now tells its investors that there will be a Braswell-based core coming branded as Pentium and Celeron in the second half of 2015. On some roadmaps leaked earlier this year Braswell was originally supposed to come in Q4 2014, then it got pushed to Q1 2015 and now probably even further. This means Bay Trail will have to last much longer than intended.
The new Skylake 14nm is going to end up as the sixth generation Core, so we would not be surprised to see them branded as the Core i7 6000 series. We are not sure how will Intel plans to have two generations coexisting in the same market at the same time, but the good news for desktop lovers is that Skylake will come to desktops and bring DDR4 to mainstream users.
AMD will face a lot of competition in 2015, but with the right pricing we believe that there is enough space for AMD-based notebooks and two-in-ones as well as desktops.
The service, which is designed to do what Drive does for Google and what Office 365 does for software rental, has gained mobile apps for the first time as Zocalo appears on the Google Play store and Apple App Store.
Amazon also mentions availability on the Kindle store, but we’re not sure about that bit. We assume it means the Amazon App Store for Fire tablet users.
The AWS blog says that the apps allow the user to “work offline, make comments, and securely share documents while you are in the air or on the go.”
A second announcement brings Zocalo into line with the AWS S3 storage on which it is built. Users will receive an update to their Zocalo sync client which will enable file capacities up to 5TB, the same maximum allowed by the Amazon S3 cloud.
To facilitate this, multi-part uploads will allow users to carry on an upload from where it was after a break, deliberate or accidental.
Zocalo was launched in July as the fight for enterprise storage productivity hots up. The service can be trialled for 30 days free of charge, offering 200GB each for up to 50 users.
Rival services from companies including the aforementioned Microsoft and Google, as well as Dropbox and Box, coupled with aggressive price cuts across the sector, have led to burgeoning wars for the hearts and minds of IT managers as Microsoft’s Office monopoly begins to wane.
nVidia has unveiled what it claims is the world’s highest-performing GPU accelerator designed for high performance computing (HPC) applications.
Launched as an addition to the Tesla Accelerated Computing Platform, the Tesla K80 dual GPU accelerator is the most powerful in Nvidia’s line-up and is aimed at accelerating a wide range of data analytics and scientific computing machine learning.
“It combines the world’s fastest GPU accelerators, the widely used CUDA parallel computing model, and a comprehensive ecosystem of software developers, software vendors, and data centre system OEMs,” said Nvidia.
The firm explained how the Tesla K80 delivers almost double the performance and double the memory bandwidth of its predecessor, the Tesla K40 GPU accelerator.
“With 10 times higher performance than today’s fastest CPU, it outperforms CPUs and competing accelerators on hundreds of complex analytics and large, computationally intensive scientific computing applications,” the firm added.
The accelerator boasts an enhanced version of Nvidia GPU Boost technology, which boosts applications by dynamically converting power headroom into the optimal performance enhancement for each individual application.
The GPU was designed to tackle “the most difficult computational challenges”, ranging from astrophysics and genomics to quantum chemistry and data analytics.
It is also optimised for deep learning tasks, a segment of the machine learning field which Nvidia says is the fastest growing.
Featuring two GPUs per board, the Tesla K80 dual-GPU accelerator doubles throughput of applications designed to take advantage of multiple GPUs.
It offers 24GB of ultra-fast GDDR5 memory (12GB of memory per GPU) as well as twice the memory of the Tesla K40 GPU, which allows the processing of double-sized datasets.
A 480GBps memory bandwidth offers an increased data throughput so that data scientists can crunch though petabytes of information in half the time compared with the Tesla K10 accelerator.
Nvidia said that the GPU’s 4,992 CUDA parallel processing cores also boost applications by up to 10 times compared with using a CPU alone.
“The Tesla K80 dual-GPU accelerators are up to 10 times faster than CPUs when enabling scientific breakthroughs in some of our key applications, and provide a low energy footprint,” said Wolfgang Nagel, director of the Centre for Information Services and HPC at Technische Universität Dresden in Germany.
“Our researchers use the available GPU resources on the Taurus supercomputer extensively to enable a more refined cancer therapy, understand cells by watching them live, and study asteroids as part of ESA’s Rosetta mission.”
The Tesla K80 dual-GPU accelerator starts shipping today from server manufacturers including Asus, Bull, Cirrascale, Cray, Dell, Gigabyte, HP, Inspur, Penguin, Quanta, Sugon, Supermicro and Tyan.
The Tesla K80 dual-GPU accelerator can also be tried for free on remotely hosted clusters.
Ubisoft is claiming that the reason that its latest Assassin’s Creed game was so bad was because of AMD and Nvidia configurations. Last week the Ubisoft was panned for releasing a game which was clearly not ready and Ubisoft originally blamed AMD for its faulty game. Now Ubisoft has amended an original forum post to include and acknowledge problems on Nvidia hardware as well.
Originally the post read “We are aware that the graphics performance of Assassin’s Creed Unity on PC may be adversely affected by certain AMD CPU and GPU configurations. This should not affect the vast majority of PC players, but rest assured that AMD and Ubisoft are continuing to work together closely to resolve the issue, and will provide more information as soon as it is available.”
However there is no equivalent Nvidia-centric post on the main forum, and no mention of the fact that if you own any Nvidia card which is not a GTX 970 or 980. What is amazing is that with the problems so widespread, Ubisoft did not see them in its own testing before sending it out to the shops. Unless they only played the game on an Nvidia GTX 970 and did not bother to test it on a console, it is inconceivable that they could not have seen it.
Amazon has become the latest vendor to commission a customized Xeon chip from Intel to meet its exact compute requirements, in this case powering new high-performance C4 virtual machine instances on the AWS cloud computing platform.
Amazon announced at the firm’s AWS re:Invent conference in Las Vegas that the latest generation of compute-optimized Amazon Elastic Compute Cloud (EC2) virtual machine instances offer up to 36 virtual CPUs and 60GB of memory.
“These instances are designed to deliver the highest level of processor performance on EC2. If you’ve got the workload, we’ve got the instance,” said AWS chief evangelist Jeff Barr, detailing the new instances on the AWS blog.
The instances are powered by a custom version of Intel’s latest Xeon E5 v3 processor family, identified by Amazon as the Xeon E5-2666 v3. This runs at a base speed of 2.9GHz, and can achieve clock speeds as high as 3.5GHz with Turbo boost.
Amazon is not the first company to commission a customized processor from Intel. Earlier this year, Oracle unveiled new Sun Server X4-4 and Sun Server X4-8 systems with a custom Xeon E7 v2 processor.
The processor is capable of dynamically switching core count, clock frequency and power consumption without the need for a system level reboot, in order to deliver an elastic compute capability that adapts to the demands of the workload.
However, these are just the vendors that have gone public; Intel claims it is delivering over 35 customized versions of the Intel Xeon E5 v3 processor family to various customers.
This is an area the chipmaker seems to be keen on pursuing, especially with companies like cloud service providers that purchase a great many chips.
“We’re really excited to be working with Amazon. Amazon’s platform is the landing zone for a lot of new software development and it’s really exciting to partner with those guys on a SKU that really meets their needs,” said Dave Hill, senior systems engineer in Intel’s Datacenter Group.
Also at AWS re:Invent, Amazon announced the Amazon EC2 Container Service, adding support for Docker on its cloud platform.
Currently available as a preview, the EC2 Container Service is designed to make it easy to run and manage distributed applications on AWS using containers.
Customers will be able to start, stop and manage thousands of containers in seconds, scaling from one container to hundreds of thousands across a managed cluster of Amazon EC2 instances, the firm said.
According to Jon Peddie Research (JPR), Nvidia has managed to claw back market share from AMD in the second quarter of 2014. JPR found that AMD’s overall unit shipments decreased 7% sequentially, while Intel and Nvidia gained 11.6% and 12.9% respectively. The ‘attach rate’ is almost flat at 155% (up 2%). A total of 32% of PCs tracked last quarter had discrete graphics, while 68% did not.
The PC market grew 6.9% sequentially, but it was down 2.6% year-on-year. Shipments of desktop graphics cards were up 7.8% from last quarter.
“Q3 2014 saw a flattening in tablet sales from the first decline in sales last quarter. The CAGR for total PC graphics from 2014 to 2017 is up to almost 3%. We expect the total shipments of graphics chips in 2017 to be 510 million units. In 2013, 454 million GPUs were shipped and the forecast for 2014 is 468 million,” JPR said.
Shipments of AMD APUs were up 10.5% over the last quarters, but AMD lost 16% in the notebook market. AMD’s discrete GPU shipments were down 19%, but notebook discrete shipments were up 10%. AMD’s overall graphics shipments were down 7%.
Intel’s desktop GPU shipments were stagnant (down 0.3%), but notebook shipments were up by 18.6%.
Nvidia’s desktop discrete shipments were up 24.3% sequentially, while notebook shipments increased 3.5% for an overall increase of 12.9%.
“Year-to-year this quarter AMD’s overall PC shipments decreased 24%, Intel increased 19%, Nvidia decreased 4%, and the others essentially are too small to measure,” the report found.
“Total discrete GPU (desktop and notebook) shipments from the last quarter increased 6.6%, and decreased 7.7% from last year. Sales of discrete GPUs fluctuate due to a variety of factors (timing, memory pricing, etc.), new product introductions, and the influence of integrated graphics. Overall, the trend for discrete GPUs has increased with a CAGR from 2014 to 2017 now of 3%.”
At the moment, an estimated 99% of all Intel chips ship with integrated graphics, compared to 66% of AMD non-server processors.
It looks like we might never see 20nm GPUs from either Nvidia or AMD. From what we know, both companies spent a lot of time looking into the new 20nm manufacturing process and they have decided that it is simply not viable for GPUs.
Yields are not where they are supposed to be and from a business perspective it doesn’t make sense to design and produce chips that would end up with very low yields. At this point we do not expect to see any high-end chips in 20nm, as there are obvious manufacturing obstacles and both companies might even skip the 20nm process altogether and move directly to 16nm FinFET.
16nm FinFET GPUs coming in 2016
We expect 16nm FinFET based GPUs sometime 2016 and this manufacturing process will bring some rather innovative products worthy of an upgrade.
One might ask why Apple doesn’t appear to have problems with its 20nm A8 and A8X chips and we might have a partial answer for you. The Apple A8 chip has to stay under 2.5W TDP, the A8X used in the iPad Air 2 A8X has a maximum TDP of 4.5W.
GPUs such as Maxwell- and Hawaii-based parts used in the Geforce GTX 980 and Radeon R9 290X have TDPs in the 150-250W range and the size of the modern GPU is an order of magnitude bigger than the size of an iPhone SoC.
Die size conundrum
The Apple A8 has a die size of 89mm2 and while we can only assume that the more powerful A8X measures over 100 mm2. Nvidia’s 28nm Maxwell GM204 die measures 398 mm2, which is about four and a half times bigger in terms of sheer die size.
To put things in perspective, in a single 20nm 300mm wafer you can place more than 700 A8 dies, while Nvidia can get about 140 Maxwell 204 chips from a 28nm High K 300 mm wafer and in 20nm manufacturing it would be able to get more, as the individual die would be significantly smaller.
However, these 150-250W chips are completely different than low-power SoCs with TDPs of less than 5W. They are worlds apart and one can assume that with the high performance and clock of discrete GPUs, coupled with their sheer size, result in higher leakage and other issues. Making a chip 4.5 times bigger means that there is much more room for potential issues, leakage and yield problems.
Don’t despair, 28nm still has some life in it
Not all is lost. We all saw that Nvidia pulled off a small miracle with the 28nm Maxwell GM204 chip, as this 5.2 billion transistor chip has a TDP of just 165W.
Its predecessor, the Geforce GTX 780 based on the GK110 chip, ended up with a 250W TDP with 7.08 billion transistors and a massive 561mm2 die size. Maxwell is also faster than Kepler, at least in this iteration, yet they are both 28nm products.
We expect that AMD’s upcoming Fiji GPU to be substantially more efficient than the Hawaii XT chip used in last year’s Radeon R9 290X. However, the new part is coming in 2015.
TSMC’s next generation 16nm process has reached an important milestone – 16nm FinFET Plus (16FF+) is now in risk production.
Needless to say, 16FF+ comes a few quarters after the 16nm rollout, expected in Q1 2015. TSMC hopes to start churning out 50,000 16FF wafers in Q2 2015. As for the Plus process, it is still more than a year away in terms of availability and it will be followed by 10nm, which is expected to materialise in late 2016.
TSMC says the improved 16FF+ process can deliver a 40% performance boost compared to its planar 20nm SoC process (20SoC), with a 50% reduction in power consumption.
“Our successful ramp-up in 20SoC has blazed a trail for 16FF and 16FF+, allowing us to rapidly offer a highly competitive technology to achieve maximum value for customers’ products,” said Mark Liu, president and Co-CEO for TSMC.
“We believe this new process can provide our customers the right balance between performance and cost so they can best meet their design requirements and time-to-market goals.”
The first 16FF+ chips are expected to tape out in late 2015 and TSMC expects the volume ramp will start in mid-2015.
AMD’s upcoming Carrizo APU has appeared in the SiSoftware Sandra and GFXBench database.
The Carrizo-L processor (AMD Eng Sample: 2M1801C1Y4381_26/18/08/04_9874 (2M 4T 2.6GHz, 1.4GHz IMC, 2x 1MB L2) is clocked at up to 2.6GHz. The new Excavator based CPU is backed by 512-core graphics and the GPU ends up almost twice as fast as previous Radeon R5 on-die solutions.
The results were unearthed by Dutch tech site Hardware.info, which pointed out that the Carrizo sample performs on a par with the A10-7300 Kaveri, despite the fact that the 7300 is clocked at 3.2GHz. This points to a significant IPC gain, as the Carrizo sample ran at 2.6GHz.
Graphics tests yielded mixed results, which may be indicative of driver issues. However, this is to be expected when it comes to upcoming parts, as the launch is still months away, giving AMD ample time to tweak the drivers. For example, the GPU core runs at 626MHz, but in different tests it was listed as a 64-bit and 128-bit part.
Red Hat has released an updated version of its Cloud Infrastructure suite that combines several products to deliver a comprehensive OpenStack-based cloud platform, adding its Satellite 6 lifecycle management tool to the mix.
Launched at the OpenStack Summit in Paris on Monday, Red Hat Cloud Infrastructure 5 brings together the firm’s Red Hat Enterprise Linux (RHEL) OpenStack Platform, CloudForms for managing hybrid cloud deployments, Red Hat Enterprise Virtualisation, and now the Satellite 6 lifecycle management tool which was released in September.
The new release is a comprehensive solution available under a single subscription licence that provides organisations with the tools they need to transform their IT infrastructure from traditional data centre virtualisation to an OpenStack-powered cloud capable of linking with public cloud OpenStack resources, Red Hat said.
“Hybrid environments are simply the reality of today’s IT, and organisations want to get to the cloud on their own terms and timeline. Red Hat Cloud Infrastructure acknowledges that reality,” said Joe Fitzgerald, vice president and general manager for cloud management at Red Hat.
“By bringing software lifecycle and configuration management capabilities that span physical, virtual and cloud systems to users via the addition of Red Hat Satellite, we’re helping to establish Red Hat Cloud Infrastructure as one of the most comprehensive and premier cloud infrastructure solutions for enterprises.”
Satellite 6 enables provisioning and lifecycle management tasks for various Red Hat products, including RHEL, while CloudForms provides cloud management and orchestration capabilities such as self-service portals, chargeback and metering of services across private and public clouds.
Meanwhile, the RHEL OpenStack Platform 5 is itself based on the previous Icehouse release of OpenStack combined with the firm’s RHEL 7 operating system. Red Hat already offers a three-year software support product lifecycle for this platform.