Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Asus did not seem to have any trouble delivering a 2.33 GHz Moorefield phone.
I am pretty sure that MSFT has got the message now, after all Ballmer and Sinofsky paid with their jobs. If with Windows 9 they can revert to a desktop orientated o/s but with optional Metro windows I believe they will get the best of both worlds and should restore their brand.
Intel is setting up new expectations in the OEMs ...
Intel Mobile and Communications unit to rebound, says Wells Fargo
http://www.theflyonthewall.com/permalinks/entry.php/INTCid2020898/INTC-Intel-Mobile-and-Communications-unit-to-rebound--says-Wells-Fargo
Wells Fargo expects the sales of Intel's Mobile and Communications unit to bottom in the March quarter, and the firm predicts that the unit's sales will grow sequentially during each quarter for the rest of the year. Wells expects the unit's losses to rise further in the June 2014 quarter and the second half of this year, and then diminish rapidly in 2015. The unit could move towards break even, potentially enabling the company to beat expectations in 2015 and 2016, added Wells, which reiterates an Outperform rating on the shares.
http://www.bloomberg.com/news/2014-06-05/intel-says-mobile-investment-needed-to-ensure-future.html
ENG is about to explode, already up 8% ! All previous resistance turned into support.
and a credible assault being made on Intel's server market share in 2016.
I will believe that when I see it because so far the ARM Server story has just been hype and false boasts.
Intel's process and design is fighting at a disadvantage against 64 bit ARM especially in the low end market.
Nonsense.
Don't you think you overreacted at the time considering the clear device progress that has now been shown in designs of all 22nm Silvermont variants ? Intel's Silvermont pricing is an ARM killer while being only a temporary nuisance to itself. The 14nm products with their better gpus can pick up the higher profit sockets later and the most significant development in 2014 is Core M not any Atom product no matter how better executed they could have been. Core M is the first very high performance fanless chip in the World and should create a new profitable market niche for Intel. There was never any reason to place so much store on smartphones however desirable a long-term goal that it is. There is plenty of time to get smartphones after eventually laying waste to ARM in tablets. Windows 9 is the next big catalyst for Intel as MSFT hopefully get the desktop/tablet hybrid o/s finally right. Then the smartphone factor can kick in.
http://www.hardwarezone.com.my/tech-news-computex-2014-intel-atom-moorefield-processor-benched-beats-snapdragon-801
Pitting reference smartphones based on the Atom Z3480 ('Merrifield') and the Atom Z3580 ('Moorefield') against the Snapdragon 801-powered GALAXY S5, the new devices were benched using 3DMark's Ice Storm Unlimited test, and the ensuing results were rather interesting.
The quad-core Z3580 scored 21,163, which beats the quad-core Snapdragon 801 (18,450), while the dual-core Z3480 trailed behind with a score of 16,350, proving that Merrifield can hold its own against the competitor's quad-core offerings. Interestingly, the physics score from the Z3580 is significant high at 21,456, compared to the Snapdragon 801 (15,866).
It's not like they are subsidizing modems, are they?
Intel is in effect subsidizing the modem subsidiary when there are low Infineon sales such as now between the crossover between 3G and 4G. Personnel still have to be paid to develop/produce the products even if very little revenue bacon is being brought home now. So there will be a double whammy improvement to mobile division's numbers when not only the contra-revenue disappears but the 4G revenue picks up. Even 3G sales will pick up again with the Sofia products.
INTC has reached year long highs which was not that difficult in the end as it already had a very favorable technical chart this year and it just needed no bad news to disrupt that. The crucial turning point was the recent dip under 25.85, if it had not rebounded above this point where 15 year old moving averages lurk it would have stayed range-bound under 26. The historical range bound nature of INTC can be explained by its chart and is not symptomatic of any fundamentals.
Cavium will be competing against the 14nm shrink so comparing Core M to say two Cavium A57-like cores as they are broadly the same 4-5W power Core M will have greater performance, performance/watt and throughput even before hyper-threading is switched on. I don't think Cavium will get any traction in the general purpose server market but it just maybe there are some niches for so many cores and their accelerators they can exploit but they will have competition from Broadcom in that respect. Power 8 remains the only serious competitive non-x86 server alternative out there at the moment although I expect Broadcom's Vulcan chip to be competitive too in some throughput respects as it uses the same one thread per issue width philosophy of Power 8. Quad-threaded, quad-issue at 3 GHz, Vulcan is the *only* serious attempt at a proper ARM server chip out there.
http://www.broadcom.com/press/release.php?id=s797235
A single A57 is about 2-3W so yes. There will also be a lot of sharing on the chip that will bring overall power down.
Inside Thunder
http://www.eetimes.com/document.asp?doc_id=1322565
Cavium plans four families of Thunder chips, targeting storage, networking, and security systems in addition to servers. Each family will use a different set of I/O and accelerator blocks from Cavium's existing silicon library.
The company plans to roll out high-end versions from all four groups initially. It aims to round out the portfolios three to six months later with 8 to 16 core versions.
The initial server SoCs will support DDR4 and integrate PCI Express Gen 3 x4 and x8 links. The server SoCs likely will pack just one or two 10G Ethernet controllers initially, with four-port 10G and 40G versions coming a year later.
Cavium will use the latest extensions of ARM's 64-bit V8 architecture, including a new exception handler (GICv3) and virtualization spec (SMMUv2). Each custom core handles out-of-order processing of up to three instructions per cycle and has a 16 Mbyte L2 cache.
In 2011 Applied Micro was the first to announce plans for a custom 64-bit ARM core. More recently AMD said it will design one, too, although it will not ship in SoCs until 2016.
Starting next year, Cavium will have a full line of ARM- and MIPS-based chips with a fair amount of overlap in the versions targeting networking systems.
"The way we see it, Thunder is more focused on cloud and data center systems and [the MIPS-based] Octeon is more embedded," says Raghib Hussain, one of Cavium's founders and its chief technology officer.
The two product lines use the same on- and off-chip interconnects, coherency protocols, and other silicon blocks. "As a result, we have the flexibility to pick and choose on a product-by-product basis whether to put in a MIPS or ARM core," says Hussain.
Nevertheless, the company now has to design two types of custom cores. "The long-term future depends on how the market evolves, but for the next generation we are committed to both" MIPS and ARM, he told us.
"I think we will see Thunder in embedded apps as well as servers," says analyst Gwennap, "so the question is over time what percent of sales will be on Thunder vs. Octeon -- over time they could shift more completely to the ARM side.
"The same thing is happening at Freescale and Broadcom -- all the major embedded guys are moving toward ARM," at the expense of MIPS and PowerPC.
— Rick Merritt, Silicon Valley Bureau Chief, EE Times
The failure of Intel to be at the forefront of the mobility explosion was primarily due to lack of vision/investment and complacency and that buck stops at the CEO. If the $7.7bn used to buy McAfee had gone to continue with XScale or even to have accelerated the current Atom roadmap by 2 years Intel would have been at the forefront of this growth market in computing rather than lagging so this was a judgement call that was not optimum in hindsight. This is not a terminal mistake as Intel is catching up now with better designs and its higher process improvement cadence.
Under the Asset Purchase Agreement, the parties have agreed that upon the consummation of the transactions, the Buyer will withhold $1,830,000 from the purchase price for the purpose of securing any purchase price adjustment to be paid by the Seller Companies in connection with any net working capital adjustment. In addition, the Buyer will deposit into escrow $6,100,000 from the purchase price (the “Indemnity Escrow Amount”) for the purpose of securing certain indemnification obligations of the Seller Companies. The Indemnity Escrow Amount will be held by an escrow agent under the terms of an Escrow Agreement until the 18-month anniversary of the consummation of the transactions, at which time the escrow agent will release the then-available Indemnity Escrow Amount to the Company (other than amounts in respect of any pending indemnification claims, which will be released upon final resolution of such claims).
“Net Working Capital Threshold” means one million, seven hundred and ninety-one thousand, one hundred and twenty-five hundred dollars ($1,791,125).
If the Estimated Net Working Capital Amount is greater than the Net Working Capital Threshold, the difference shall be added to the Purchase Price paid to the Seller Companies at Closing pursuant to Section 2.01(k), and if the Estimated Net Working Capital Amount is less than Net Working Capital Threshold, the difference shall reduce the Purchase Price paid to the Seller Companies at Closing pursuant to Section 2.01(k). The Purchase Price to be paid to ASTC at Closing pursuant to Section 2.01(k) as adjusted pursuant to the foregoing sentence shall constitute the “Estimated Adjusted Purchase Price.”
Current net working capital as of March 31, 2014 not including the long term debt of $5.8m was $1.9m so in effect LMT are buying ASO as of the end of the last quarter.
ASTROTECH CORPORATION TO SELL ASTROTECH SPACE OPERATIONS BUSINESS TO LOCKHEED MARTIN FOR $61 MILLION - SEC FILINGS
http://www.sec.gov/Archives/edgar/data/1001907/000157104914002183/t1401006_ex99-1.htm
http://www.sec.gov/Archives/edgar/data/1001907/000157104914002183/t1401006_8k.htm
CONTRACT AGREEMENT
http://www.sec.gov/Archives/edgar/data/1001907/000157104914002183/t1401006_ex2-1.htm
Chart will hunt down the shorts over time. The first non-ASO quarter won't be reported until about a year from now, plenty of time for the chart and 1stDetect to work its magic.
Check ENG out too, great chart just like ASTC, above all its moving averages going back years. Not a bad little company too, just announced two consecutive profitable quarters after cutting the fat. Any volume will send it soaring over 4 like it was only a few days ago.
I assume that some of that can be offset against losses over the last few years.
They would not sell both ARM and Intel SOC's to the same customer.
Why not especially when their ARM product does not contain an integrated modem. A Silvermont quad-core would also be above the performance of their current top of the line quad-core 1.8 GHz A12 chip.
http://en.wikipedia.org/wiki/Rockchip#RK32xx_series
I could even see Rockchip going all x86 from here with Airmont at 14nm and Goldmont at 10nm. Intel would of course pick up good foundry margin then.
Intel is actually designing the SOC with Rockchip guidance.
I believe it's the other way around. This is a way for x86 to become a Chinese native product which will obviously help penetration.
Rockchip implications ... Intel is basically acting like ARM in this deal and licensing its IP to Rockchip so there will not be much money in this deal, maybe a dollar or less a chip for 4 cores and the modem. Where this gets interesting is in the future, on 14nm.
I can see this exercise being repeated but on Intel's 14nm so Intel would pick up foundry margin as well then. The deal for Intel is that it will open the door to China as it still will be a natively designed SoC using licensed IP (by Intel not ARM) and built on an external foundry (by Intel not TSMC) as Rockchip do now. Not a bad localized deal by BK with good future implications.
Ok reading it that way is not so bad except that Moorefield is shown no earlier than Q414 and the phone version of Goldmont, Morganfield is shown as Q415 as opposed to Q215 for the Willow Trail tablet version.
AP Smartphone Roadmap
http://s1.directupload.net/images/140528/tngisidw.png
Android Smartphone FFRD Roadmap
http://s14.directupload.net/images/140528/ziuz6teq.png
AP Tablet Roadmap
http://s7.directupload.net/images/140528/e58x2mb4.png
Tablet FFRD Roadmap
http://s1.directupload.net/images/140528/c3lb7cvd.png
AP Roadmap Overview
http://s7.directupload.net/images/140528/py2xpry4.png
It shows there was not much wrong with the design roadmap just its execution.
By their own estimates Merrifield is now 1.5 years late and Moorefield/Morganfield(Broxton) 1 year late. They all would have been ultra-competitive if released on time.
Goldmont is single-threaded from those slides so still extra performance being left on the table. Also tops out at 2.7 GHz. Good detail on those slides, nice work.
Intel extends incentives to boost development of Intel-based tablets
http://www.digitimes.com/news/a20140522PD213.html
Monica Chen, Taipei; Adam Hwang, DIGITIMES [Thursday 22 May 2014]
Intel originally offered incentives for selected China-based white-box vendors such as Ramos to adopt Intel platforms for tablets, and has extended to additional China-based white-box vendors and international tablet vendors, in an attempt to reach total shipments of 40 million Intel platform-based tablets in 2014, according to Taiwan-based supply chain makers.
More than 20 China-based white-box tablet vendors have adopted Intel platforms, the sources said. In addition, Asustek Computer, Acer, and Lenovo as well as Hewlett-Packard and Toshiba have developed or will develop Android and/or Windows 8 tablets equipped with Intel processors, the sources indicated. With about 30 supporting vendors, Intel is expected to reach its target shipments of 40 million tablets, the sources noted.
Intel incentives include very low quotes for processors and marketing subsidies, the sources noted. Viewing that Apple has dominated the global market for high-end tablets, and tablet markets in the US and Europe have reached saturation, Intel has focused on promoting tablet platforms on China where demand for inexpensive models is growing fast, the sources said.
For Intel, the offering incentives is equivalent to giving up profitability to pursue global market share, the sources said.
http://www.dailytech.com/Chinese+Government+Says+Windows+8+is+too+Expensive+Will+Cling+to+XP/article34781.htm
http://www.dailytech.com/Chinese+Government+Bans+Windows+8+for+Security+Reasons/article34932.htm
China has said that it plans to continue to use Windows XP on most government computers. The government says it will patch the aging operating system itself, rather than pay Microsoft large fees as the UK government has agreed to.
The other deal about reference tablets is that they are deliberately tuned for performance e.g. Tegra 4 which no Tegra 4 production tablet has come close to matching. I think you will have to wait for the 20nm shrink of this to see a genuine AMD long life fanless tablet processor but still it appears a good improvement for AMD still at 28nm, keeps them in the game and AMD investors should keep a close eye on it for any sales progress as benefits could be material.
The Pro versions of Microsoft’s device are outshipping the RT version, which uses ARM-based chips, by a margin of about four to one, according to IDC.
Microsoft and Qualcomm have been working together on Surface for about a year and the project had been proceeding until recently, when Nadella and Elop declined to give it the final go ahead, the people said.
As a result, Microsoft doesn’t currently have an approved project to create a smaller tablet but it could look at the category again in the future, said one of the people.
AMD would not leave it long it enough with any reviewer for them to run any battery tests which is kind of telling and the chassis was described as bulky and AMD need this kind of improvement anyway not to be blown away by cheaper longer lasting Bay Trails and more potent Cherry Trails. Other than that it reminds of Nvidia's 28nm K1 with its strong gpu also which unlike Mullins already has got a good design win at Xiaomi in a retina 8" tablet. There is no point in it being a 'rout' as you think if no-one builds a design around it.
http://www.bloomberg.com/news/2014-05-20/microsoft-said-to-back-off-plans-to-debut-smaller-surface.html
Microsoft Corp. (MSFT), struggling to break into the tablet market, backed off plans to introduce a smaller version of its Surface tablet based on Qualcomm Inc. (QCOM) processors, people with knowledge of the decision said.
Microsoft Chief Executive Officer Satya Nadella and Executive Vice President Stephen Elop decided that the product in development wasn’t different enough from rivals and probably wouldn’t be a hit, said one of the people, who asked not to be identified because the plans weren’t public. Engineers had been working on the device and had planned to unveil it as early as today at an event in New York, two of the people said.
Instead, the company removed any references to Surface devices based on Qualcomm’s processors from the event and unveiled Surface Pro 3, an upgraded version of its pricier Windows-based tablet, which uses chips from Intel Corp. (INTC) and features a larger 12-inch screen. Microsoft is still working on smaller versions, Panos Panay, who runs the Surface business, said in an interview after the event. Panay declined to provide details on future products.
Microsoft’s Surface had just 1.3 percent share of the market in the first quarter, while smaller-size tablets accounted for 53 percent of sales in the period, according to researcher IDC. Nadella has said he wants to turn the Redmond, Washington-based company into one that focuses on “mobile first, cloud first.” ...
It won't have in 64-bit, in fact it will be leading the 64-bit Android hardware/software about a year ahead of the ARMy and these will be the premium highest margin products of their day. Merrifield and Moorefield will be the first early wave.