There have been few details released on the new 10 core Skylake-X which will replace the Intel Core i7-6950X Extreme Edition which has just been released.
Intel has just released its Broadwell-E generation of ten, eight and six cores with Intel Core i7-6950X Extreme Edition being the fastest end the most expensive. But we have managed to get a few details about its replacement – the Skylake-X.
You can expect two SKUs, 140W X versions with 10 cores and the one will less cores called the K version. The new Extreme edition CPUs will have the new R4 socket. This new socket is also called LGA2066 some 55 higher than with the existing socket number.
There will be a Kaby Lake-X 4 core processor with 95+ W TDP using the same LGA 2066 R4 socket . Both Skylake-X and Kaby Lake-X support the new chipset that is known as Kaby Lake-X.
This is 200 series chipset will be the successor of the Skylake 100. The new chipset will come with up to 24 PCIe 3.0 PCI express lanes. In fact this is the only major difference in the chipset. It does support Octane storage techhonlogy, something that 100 series chipset cannot. The Kaby Lake 200 series chipset supports 6 SATA 3.0 ports, up to 10 USB 3.0 ports, DMI 3.0, up to three 4Xports for PCIe 3.0 drives.
If the Zen desktop core gives Intel some serious competition, we bet that Intel won’t charge $1700 for its highest end overclockable desktop CPU. Zen is still at least few months away, we expect it at late 2016 at best.
Intel said that servers with the Intel Xeon E5-4600 v4 family can now have up to 22 cores and 44 threads for enough processing power for most scale-out and large workloads.
Intel’s new range is claimed to be 2.6x better than previous generations Along with the performance boost and the higher core and thread count, the new E5-4600 v4 family can provide up to 55MB of last-level cache, support up to 6TB of DDR4 2400 memory, and 40 lanes of PCIe 3.0.
The new processors have AES encryption and fast public key (RSA) encryption along with a strong random number generation that enables hardened, pervasive data protection without impact to application response times. The family is supplied with Intel’s Intelligent Power technology to improve power across both the CPU and memory. The latest version of processors family supports per-core P states (PCPS) to optimize the power usage of each processor core.
The new family has what Chipzilla calls advanced multi-core, multi-threaded processing – up to 22 cores (previously up to 18) and 44 threads (previously up to 36) per socket for running more and heavier workloads and higher density of VMs per server.
There is a larger cache: Up to 55MB (Previously up to 45MB) of last level cache for fast access to frequently used data. There is a faster memory with up to 48 DIMMS per four-socket server for memory-intensive applications and faster maximum memory speeds with DDR4 memory. Chipzilla claims that this gives higher performance for demanding workloads.
It has Optimized Intel Advanced Vector Extensions 2.0 (Intel AVX 2.0) enables applications to run at maximum “turbo” frequencies wherever possibl and IIntel Turbo Boost Technology 2.0 acceleration takes advantage of power and thermal headroom.
Flexible, high-performance hardware-enhanced virtualization: Improve overall reliability and responsiveness through new Intel Virtualization Technology features, including New Posted Interrupts, Page Modi cation Logging, and VM Enter/Exit latency reduction
It also has multiple rank sparing DDR4 recovery for command and address parity errors and the latest Intel Data Protection Technology.
Intel tells us that the new Intel Xeon E5-4600 v4 processor family is available now.
In what it calls a drive further into the data centre market. Cavium has entered into a definitive agreement to acquire all outstanding QLogic common stock in a deal worth approximately $1.36 billion.
The acquisition adds Qlogic’s Fibre Channel and Ethernet controllers and boards to Cavium’s line up of communications, security and general-purpose processors making Cavium a full-line supplier to data centres.
It also means that Cavium can take on storage and networking with Broadcom, Intel and Mellanox. The deal also gives Cavium a mature software stack in storage and networking and operational savings expected to amount to $45 million a year by the end of 2017.
Both companies sell to server makers and large data centers with a customer overlap of more than 60 per cent. Qlogic’s customer base is highly concentrated with nearly 60 per cent of its business for the last several years to HP, Dell and IBM.
Researchers at the University of California, Davis, Department of Electrical and Computer Engineering have developed 1000-core processor which will eventually be put onto the commercial market.
The team, from t developed the energy-efficient 621 million transistor “KiloCore” chip so that it could manage 1.78 trillion instructions per second and since the project has IBM’s backing it could end up in the shops soon.
Team leader Bevan Baas, professor of electrical and computer engineering said that it could be the world’s first 1,000-processor chip and it is the highest clock-rate processor ever designed in a university.
While other multiple-processor chips have been created, none exceed about 300 processors. Most of those were created for research purposes and few are sold commercially. IBM, using its 32 nm CMOS technology, fabricated the KiloCore chip and could make a production run if required.
Because each processor is independently clocked, it can shut itself down to further save energy when not needed, said graduate student Brent Bohnenstiehl, who developed the principal architecture. Cores operate at an average maximum clock frequency of 1.78 GHz, and they transfer data directly to each other rather than using a pooled memory area that can become a bottleneck for data.
The 1,000 processors can execute 115 billion instructions per second while dissipating only 0.7 Watts which mean it can be powered by a single AA battery. The KiloCore chip executes instructions more than 100 times more efficiently than a modern laptop processor.
The processor is already adapted for wireless coding/decoding, video processing, encryption, and others involving large amounts of parallel data such as scientific data applications and datacentre work.
In an official slides that have leaked, AMD has confirmed most of the specifications for both the Polaris 10 and the Polaris 11 GPUs which will power the upcoming Radeon RX 480, RX 470 and RX 460 graphics cards.
According to the slides published by Computerbase.de, both GPUs are based on AMD’s 4th generation Graphics Core Next (GCN 4.0) GPU architecture, offer 2.8 perf/watt improvement compared to the previous generation, have 4K encode and decode capabilities as well as bring DisplayPort 1.3/1.4 and HDR support.
Powering three different graphics cards, these two GPUs will cover different market segments, so the Polaris 10, codename Ellesmere, will be powering both the Radeon RX 480, meant for affordable VR and 1440p gaming as well as the recently unveiled RX 470, meant to cover the 1080p gaming segment. The Polaris 10 packs 36 Compute Units (CUs) so it should end up with 2304 Stream Processors. Both the RX 480 and RX 470 should be coming with 4GB or 8GB of GDDR5 memory, paired up with a 256-bit memory interface. The Ellesmere GPU offers over 5 TFLOPs of compute performance and should peak at 150W.
The Radeon RX 470 should be based on Ellesmere Pro GPU and will probably end up with both lower clocks as well as less Stream Processors and according to our sources close to the company, should launch with a US $179 price tag, while the RX 480 should launch on 29th of June with a US $199 price tag for a reference 4GB version. Most AIB partners will come up with a custom 8GB graphics cards which should probably launch at US $279+.
The Polaris 11 GPU, codename Baffin, will have 16 CUs and should end up with 1024 Stream Processors. The recently unveiled Radeon RX 460 based on this GPU should come with 4GB of GDDR5 memory paired up with a 128-bit memory interface. The Radeon RX 460 targets casual and MOBA gamers and should provide decent competition to the Geforce GTX 950 as both have a TDP of below 75W and do not need additional PCIe power connectors.
According to earlier leaked benchmarks, AMD’s Polaris architecture packs quite a punch considering both its price and TDP so AMD just might have a chance to get a much needed rebound in the market share.
AMD has released a short video where its lead system engineer Louis Castro running Doom on its Summit Ridge, Zen-based processor.
This means that the silicon is in good shape and the processor was taped our probably late last year with no major issues. AMD’s CEO Lisa Su has already said that the desktop version shall arrive first, and this was the CPU demonstrated in the video.
Summit Ridge is not an APU and doesn’t have a GPU core. AMD engineers were using a discreet GPU probably from one they found out the back.
The Summit Ridge is an FM4 socket processor and half dozen of them are shown in the video.
AMD’s Zen chip will have as much as 32 cores, 64 threads and more L3 cache than you can poke a stick at.
Codenamed Naples, the chip uses the Zen architecture. Each Zen core has its own dedicated 512kb cache. A cluster [shurely that should be cloister.ed] of Zen cores shares a 8MB L3 cache which makes the total amount of L3 shared cache 64MB. This is a big chip and of course there will be a 16 core variant.
This will be a 14nm FinFET product manufactured in GlobalFoundries and supporting the X86 instruction set. Naples has eight independent memory channels and up to 128 lanes of gen 3 PCIe. This makes it suitable for fast NVMO memory controllers and drives. Naples also support up to 32 SATA or NVME drives.
If you like the fast network interface, Naples supports 16x10GbE and the controller is integrated, probably in the chipset. Naples is using SP3 LGA server socket.
The first Zen based server / enterprise products will range between a modest 35W TDP to a maximum of 180W TDP for the fastest ones.
There will be dual, quad, sixteen and thirty-two core server versions of Zen, arriving at different times. Most of them will launch in 2017 with a possibility of very late 2016 introduction.
It is another one of those Fudzilla told you so moments. We have already revealed a few Zen based products last year. The Zen chip with Greenland / Vega HBM2 powered GPU with HSA support will come too, but much later.
Lisa Su, AMD’s CEO told Fudzilla that the desktop version will come first, followed by server, notebook and finally embedded. If that 40 percent IPC happens to be across more than just a single task, AMD has a chance of giving Intel a run for its money.
The rumor mill is suggesting that Intel is about to kill off its Atom’s for server program.
In 2013, Intel started work on a server processor code-named Avoton and a networking-oriented counterpart called Rangeley which were based on Silvermont. Silvermont was a low-power, low-cost Atom processor which was aimed at micro-servers.
Intel’s roadmap was to follow Avoton with Denverton, using next-generation Goldmont CPU cores and built on the company’s 14-nanometer chip manufacturing technology. But Intel keeps making excuses for its delay or pointing into the horizon and saying “oh look there is a badger with a handgun.”
First, it was supposed to be here in 2016 and later Intel roadmaps said 2016. The latest one says “beyond 2016.” Duke Nukem was more specific, and so was Zen.
Intel’s higher-performance Xeon processors will be a better proposition by the time this one comes out and face it who is going to want a chip that should have come out in 2015 in 2017.
Logically Intel must be thinking of giving up on it. The reason could be that few people are interested in Avoton and Rangeley and this could mean an end of Atom-based server chips as a concept.
Intel must be thinking that the time wasted on Atom-based server chips would better spent plugging its Xeon D product line.
Intel seems to be slimming down a lot lately. This week it let its first batch of workers go from its Irish plant. The last thing it wants is pushing chips that no one wants. Of course this leaves the area open to Arm, but so far no one has demonstrated that is an option either.
Intel is gearing up to launch its third generation 14nm processor codenamed Kaby Lake which is a minor update to the Skylake generation.
This quad-core with eight threads is a high-end processor that will bring a few small changes. It will be faster within the same TDP, but Intel is not talking about the actual clocks. The fastest Skylake is the Intel Core i7-6700K and it has a 4.0GHz base clock that reaches 4.2GHz on Turbo.
Kaby Lake Core i7 7x00K series will be slightly faster. Intel is telling its partners that the Kaby Lake-S will have 15 percent more high speed input / output lanes. This is only four additional PCI express lanes and nothing that is going to change the world.
Intel Rapid Storage 15 is unique to Kaby Lake-S and the Intel Optane storage technology support. Intel has made us update motherboards to support PCI-e drives and it will do the same with Intel Optane, storage technology that will rely on PCI Express Gen 3.
Intel supplies DDR4-2400 memory support and it plans to teach its graphics a few new tricks. The desktop CPUs are expected in Q4 2016 which is quite late to the game, but better late than never.
The upgrade market will buy it as long as it arrives before Thanksgiving and Christmas.
Intel has not given up on its power saving roadmap even if it will not produce any more phone SoCs. In fact its Apollo Lake notebook platform can bring up to 30 percent better performance.
This chipset is targeted at tablets, 2-in-1 tablets, small desktops, Chromebooks and Cloudbooks – cheap Windows based notebooks to counter Google’s Chromebooks.
Apollo Lake brings 30 percent better CPU performance mostly due to the better graphics performance which arrive with the updated Intel Gen9 graphics. The platform supports DDR3L, LPDDR3, and LPDDR4 memory giving OEM’s more options and the platform should have the 15 percent longer battery life.
There is an official support for USB Type-C but we were under the impression that Braswell-based notebooks / tablets support the new USB too.
A leaked slide indicates that Apollo Lake could end up branded as Intel Pentium and Celeron N/J series. The top of the line version will replace the Intel Pentium N3700 and will have 2MB cache and 2.4GHz clock quad core. It is expected in the second half of 2016.
However, Intel is putting the emphasis on analytics with this generation, claiming to offer greater performance and reliability for crunching data.
Available immediately in systems from enterprise vendors including HPE, Dell, Lenovo, SGI and Fujitsu, the updated Xeon E7 family comprises the E7-8800 v4 chips aimed at eight-socket servers and the E7-4800 v4 chips aimed at four-socket servers.
Formerly known as Broadwell-EX, the new chips deliver the usual boost in performance over the previous generation, especially with an increase in the number of maximum cores from 18 to 24, and a doubling of the maximum supported memory to support up to 24TB per system.
This latter capability is to enable new systems based on the Xeon E7 family to meet the growing demands of enterprise applications such as in-memory databases and analytics involving large datasets. In fact, the Xeon E7 chips show a doubling of performance for processing analytic queries, according to Intel, as measured by TPC-H benchmark results.
The emphasis on analytics is because the field is becoming a ubiquitous part of almost every business application, according to Nidhi Chappell, product line manager at Intel’s Data Centre Group.
“It is fair to say that analytics is pretty ubiquitous. There has a been a lot of research that has shown that companies that use advanced analytics tend to have more data-driven decisions, are able to make better insights and generally enjoy better competitive advantages,” she said.
However, there are barriers to wider adoption of analytics, according to Intel, including that the software stacks to deliver a solution are complicated and call for specialised skills. It can be hard to identify the right data to analyse, and it can be difficult to get actionable insights at the end of the process.
The firm is trying to address these with the Trusted Analytics Platform (TAP) initiative, which is described as an open source effort aimed at providing the tools and services with which IT departments, data scientists and application developers can collaborate to link big data with applications.
As well as improved performance, Intel said that TAP has enhanced reliability for mission-critical applications, adding new capabilities to its Run Sure technology that includes an updated version of Intel’s Machine Check Architecture Recovery mechanism and improvements to virtual machine handling such as lower latency when entering and exiting virtual machines.
Interestingly, Intel also claimed performance gains in the Xeon E7 v4 family against IBM’s Power8 processors, which suggests that Intel now sees the Power architecture as serious competition in the data centre market.
Intel claimed that the top-end E7-8890 v4 chip has 1.4x the performance of Power8, with half the operational costs for an eight-socket system, and up to 10 times the performance per dollar.
The claims are notable in that IBM has said previously that Power chips offer much better performance per dollar than x86 chips, partly by being able to handle more threads per core.
As with the previous generation of Xeon E7 chips, enterprise vendors can combine the processors with node controller hardware to scale beyond eight sockets and up to 64 sockets.
MediaTek’s R&D teams are working with European-based car vendors to develop the company’s automotive electronics and virtual reality (VR) offerings.
Digitimes claims that having developed SoCs for smartphones, mobile devices, and connected home appliances, MediaTek is stepping up development of chips solutions for auto electronics and VR applications.
MediaTek is focused on in-car entertainment systems, and will be using its partnership with China-based NavInfo, a digital mapping service provider to help out.
NavInfo will sell subsidiary AutoChips (Hefei) and will also form a strategic alliance in which MediaTek will make an investment of US$100 million.
MediaTek will be developing VR for handsets and will support Google’s Daydream VR platform.
Meanwhile the team is flat out improving its IC solutions for Internet of Things (IoT) and wearable device applications. It is pretty sure that this will become the third largest segment after mobile devices and connected home appliances such as digital TVs. In fact the only two areas that MediaTek does not appear interested in is server and augmented reality (AR) applications.
The Micron 1100 SSD is a more mainstream oriented SSD that will be based on Marvell’s 88SS1074 controller and Micron’s 384Gb 32-layer TLC NAND. Using a SATA 6Gbps interface and available in M.2 and 2.5-inch form-factors, the Micron 1100 should replace Micron’s mainstream M600 series, based on 16nm MLC NAND.
The Micron 1100 SSD will be available in 256GB, 512GB, 1TB and 2TB capacities. It will offer sequential performance of up to 530MB/s for read and up to 500MB/s for write with random 4K performance of up to 92K for read and up to 83K IOPS for write. With such performance, it is obvious that the Micron 1100 series will target mainstream market and be a budget SSD.
The Micron 2100 is an M.2 PCIe NVMe SSD that is actually Micron’s first client oriented PCIe SSD and also the first PCIe SSD based on 3D NAND. Unfortuantely, Micron did not finalize the precise specifications so we still do not have precise performance numbers but it will be available in capacities reaching 1TB.
The Micron 1100 is expected to hit mass production in July so we should expect some of the first drives by the end of the next month. The Micron 2100 will be coming by the end of summer.
Nvidia might have scored a few wins by touting its GPU’s in the HPC market, but it is starting to lose ground to the co-processor, according to Intel’s Diane Bryant.
In an IDC interview Intel’s data center boss said that Nvidia gained an early lead in the market for accelerated HPC workloads when it positioned its GPUs for that task several years ago. However there is a perception that processors used for machine learning today are GPUs like those from Nvidia and AMD.
Bryant was a bit miffed when she was asked how Intel can compete in this market without a GPU. She said that the general purpose GPU, or GPGPU was just another type of accelerator and not one that’s uniquely suited to machine learning.
It is better to look at Knights Landing which is a coprocessor, but it’s an accelerator for floating point operations, and that’s what a GPGPU too.
She said that since the release of the first Xeon Phi in 2014, Intel now clawed back 33 percent of the market for HPC workloads that use a floating point accelerator.
“So we’ve won share against Nvidia, and we’ll continue to win share,” she said.
She said that Intel’s share of the machine learning business may be much smaller, but the market is still young.
“Less than one percent of all the servers that shipped last year were applied to machine learning, so to hear Nvidia is beating us in a market that barely exists yet makes me a little crazy,” she says.
Intel will continue to evolve Xeon Phi to make it better at machine learning tasks. She said that there are two aspects to machine learning – training the algorithmic models, and applying those models to the real world in front-end applications. Intel’s FPGAs and its Xeon processors mean Intel has both sides of the equation covered.
But Nvidia’s GPUs are harder for programmers to work with which could give Intel an edge as ordinary businesses need to adopt machine learning. Knights Landing is “self-booting,” which means customers don’t need to pair it with a regular Xeon to boot an OS.
However Intel’s newest Xeon Phi has a floating point performance of about 3 teraflops, which is a little slow compared to the five teraflops for Nvidia’s new GP100.
Intel’s restructuring axe seems to be falling on its PC client division and software areas with more than 12000 jobs to go.
Our well-placed sources are confident that the PC group will be the hardest hit. This is all because the PC market has stopped growing and Intel has to find its way to new markets to supplement loss of this business.
Latest research data from IDC indicates that in 2016 PC market will decline from 275.8 million units in 2015 to 260.8 million units in 2016 and the current projections for 2017 show the PC market slightly decreasing to 257.9 million units. At its peak PC market was at 364.0 million units, but this was in 2011 when things were rosier, kids were polite to their parents, and rock stars played decent music. These times are clearly behind us and Intel knows it.
The PC group downsize is being supervised by Dr. Venkata “Murthy” Renduchintala who is Intel’s number two. He is the bloke who was paid $25 million dollars to defect from Qualcomm. Murthy has already done a high level clean up at PC client group and is believed to be thinking about dusting the top of the corporate bookshelf next.
Another team which will be pummeled is Rene James’s old software outfit. People from software services and the security division formerly known as McAfee are expected to mostly go the same way as the artist formerly known as Prince.
Murthy’s also wants to get Intel to the right course with IoT market. Marketing for that area is expected to grow from $655.8 billion in 2014 to $1.7 trillion in 2020. Intel wants the piece of that cake, and perhaps a few tea and biscuits to go with it and it will be interesting to look the fight in this promising land market.
There is still no killer app to help the IoT market which defines it. IoT right now is nothing and everything.