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.
I was thinking more along the lines of those using PowerVR or Mali at 28nm and Moorefield is aimed square at the 800 series.
Heh, that's what experience and advanced years buys you .* Anywayyy, this now verifies that Merrifield is flexible enough to be paired with any modem going forward which should be good for future socket longevity especially as that gpu is not going to be easily beat by a 28nm ARM foundry part.
* I really did not mean those previous remarks as a personal sleight just as a matter of fact statement. You will gain both maturity and experience over time.
how much "Contra Revenue" do you think Merrifield costs Intel? (per chip)
Absolutely nothing, it's a ~$5 cost chip and every one initially sold should be at a profit. Eventually in its declining years it could be sold around cost for market share reasons like Medfield and CloverTrail+. However its powerful gpu and 64-bitness should ensure future longevity.
You should ask your contacts now for clarification but I suspect it is a marketing slide to show the most likely configuration rather than the only one.
It's a marketing choice not a technical necessity
http://newsroom.intel.com/community/intel_newsroom/blog/2014/02/24/intel-gaining-in-mobile-and-accelerating-internet-of-things
Moorefield is optimized for Intel's 2014 LTE platform, the Intel® XMM™ 7260, which the company also introduced today.
i.e. optimized to give us the most money in a phone package .
If 7260 can be paired with ARM chips it sure as hell can be paired up with Merrifield if an OEM so chooses. That Intel slide was meant more to show as to what Merrifield and Moorefield can be paired up with at launch and Anand even contradicts that in his article
http://www.anandtech.com/show/7789/intel-talks-merrifield-moorefield-and-lte-at-mwc-2014
On the modem side, Intel hopes to pair Merrifield with the XMM 7260 LTE modem.
Heh, so all the old PC OEMs will be Intel's mobility division. It is time for Dell to show everyone how to build the cheapest phone while still making money for him and Intel .
Good, then that ability plus its 64-bitness should ensure future sales for this chip in the years to come.
but given that Merrifield is stuck with XMM 7160
Why is it stuck with that particular discrete modem ?
You have had zero work experience dealing with computers in a business environment so what I said about you was a fact not an insult. Also the chart you displayed was a tablet one not a phone one. Perhaps you should also think about the insults you are so liberally spreading about a fine engineering product which frankly are unwarranted. You should also think about your original unrealistic expectations in encroaching quickly on an incumbent's market and Intel has been a whole lot successful so far in getting tablet share than ARM has been in getting PC share. Phones will take longer as performance at least is not a big issue for customers so OEMs can choose to ignore Intel if they see fit.
They know Merrifield is a dud
Only in your immature inexperienced opinion.
Intel is learning, but they are doing a LOT of damage to their credibility by claiming that these chips are super duper rather than admitting the truth that they're late and not as fully integrated/featured as the market wants.
No, by highlighting their chip strengths they are embarrassing all the current big names out there. It is making a monkey out of both the S800 and your previous favorite A7 so far. The only one who is in damage control is you having been shown all your previous technical fears were unfounded.
Krait's poor estimated SpecINT performance should have given someone a clue it is the weakest of the current 3-issue ARMy. What's more important is why the much vaunted Apple A7 could not beat a 2.13 GHz Silvermont at it either .
Gotta love Imagination Tech.
Yes but a big part of your Merrifield concerns pre-launch was that they would put an inferior PowerVR6 gpu variant to the one in A7 when in fact they have done the complete opposite and put a better one in.
To me, Merrifield looks an excellent modern phone chip fit for purpose. Everything about it says that Intel have balanced correctly the conflicting requirements of phone form-factor cpu performance, gpu performance, power consumption into the smallest die possible at 22nm. The engineers have done their job well. I certainly would have no qualms about making this my first smartphone purchase eventually as technically it ticks all the boxes for me and I know there will no performance/power/heat surprises under load unlike all the other quad-cores out there.
Now you may argue that's not what is required for marketing purposes but Intel does not have a tinier phone chip core than Atom (Quark too under-powered) to make a cheap quad-core and I also believe there will be enough discerning customers out there eventually to correctly ascertain the technical superiority of a powerful dual-core (e.g. Merrifield, A7, Denver) against a weaker quad-core (e.g. A7, A9, A53) in a ~2W limited phone form factor. Moorefield will address these marketing concerns shortly anyway when timed with XMM7260 release it can achieve the maximum impact.
So where's all the design wins you shout ? Well, unlike you I never believed established mobility OEMs would rush to put Intel chips into their designs even if they were say 20-30% better. Why would they rush to increase Intel's strength in mobility like they have in PCs and Servers ? No, Intel will have to win mobility from the bottom up using customer power and using established PC OEMs like Asus and Lenovo wanting to become major new mobility players too. Intel has given them a good phone chip in Merrifield, now it is up to them to design a good phone chassis around it that highlights its performance/power ability over the competition.
Finally when your Broxton hero chips are released in 2015 Intel will still need a very cheap phone/tablet chip to put into sub $200 devices and it is better then for margin/cost reasons that it is a purpose built dual-core like Merrifield rather than a disabled larger die Moorefield quad-core.
Perhaps tablets demand quad core?
Probably in 10"+ models but as a Medfield/CloverTrail+ replacement in cheap tablets/phones it looks good to go when they are both eventually pensioned off.
Who says they won't eventually. Seems a good fit. If CLT+ can do it ...
Educated guess but I suspect like Medfield its die size will start with a 6 and probably cost only that much in dollars to make.
But in 1H 2014? It's decidedly less impressive.
A7 performance though in 2/3 the die space, I suspect Intel can churn them out at around $5 cost. It is obvious Intel built a mobile chip that hit all the phone performance/power/price objectives but perhaps not the crazy marketing objectives needed today. No matter, Moorefield will be along for that. Merrifield just needs someone to put a nice slim chassis around it to highlight its efficiency.
That was reasonably fair and balanced apart from this line 'Merrifield isn't quite good enough to go up against a Snapdragon 800 (from either a raw performance standpoint' in which you should have stipulated that as multi-thread/task performance as I suspect in real-world usage Merrifield will win more than it loses against S800 unless you are really taxing your phone in multi-tasks/threads and even then when throttling kicks in the difference could still be negligible.
Apple A7 was the phone chip you decided was state of the art when it came out, well according to Intel's benches it more than holds its own against it. They also did put a decent gpu on it after all .
it's not going to get a lot of traction.
You don't know that in terms of eventual sales. I think it is an excellent small efficient chip that will do a lot of disrupting when discerning buyers actually bother to compare delivered results rather than specs. It will be a thorn in the ARMy's side for years to come when it cascades lower down the food chain. Intel have built 3 phone chips and not one of them has been bad for its intended purpose, the x86 phone chip brand has got off to a good start. Servers started inauspiciously too with the lowly Pentium Pro but the brand just kept getting better and better.
Merrifield performs better than I expected, notice it is beating the heavily touted Apple A7 in cpu and gpu as well as making a monkey of S800 in performance and battery life. Hats off to the Intel engineers for building the right technical chip for the form factor. Discerning buyers will notice this too.
Noticeably quicker/smoother than the 32nm x86 phone chips. Just has to now show battery life differentiation as well to get users to really notice it from the ARMy crowd.
When is 28 nanometers faster than 22?
Kaveri is baked in a 28-nanometer, planar, bulk silicon process, which is nowhere near as efficient as state-of-the-art FinFET (what Intel calls "Tri-Gate") or even the less-than-TriGate, more-than-bulk – and somewhat expensive – silicon-on-insulator (SOI) process that was used in Kaveri's predecessor.
There were reasons to go with 28nm rather than 22nm, Macri told us, that were discovered during the design process. That process was run by what he identified as a "cross-functional team" composed of "CPU guys, graphics guys, mixed-signal folks, our process team, the backend, layout team."
That cross-functional crew identified a boatload of process variants, and members of the team each ran tests based on their areas of interest, examining such factors as power curves and die-area needs.
"What we found was with the CPU with planar transistors, when we went from 28 to 22, we actually started to slow down," he said, "because the pitch of the transistor had to become much finer, and basically we couldn't get as much oomph through the transistor."
The problem, he said, was that "our IDsat was unpleasant" at 22nm, referring to gate drain saturation current*. In addition, the chip's metal system needed to be scaled down to fit within the 22nm process, which increased resistance.
"So what we saw was the frequency just fall off the cliff," he said. "This is why it's so important to get to FinFET."
http://www.theregister.co.uk/2014/01/14/amd_unveils_kaveri_hsa_enabled_apu/?page=3
That cloud has magical powers and it is rising !
when it costs you more to produce a product or service than you get in revenue, that's pretty much a disaster
You are misunderstanding the GM of the payload business. The first million or so of cost is the fixed costs involved in running two big spaceports. Breakeven is around 2 missions, good profit is at 3 and loss is at 1. The GMs of the actual missions themselves are always quite high, ~50%.
It's possible that Samsung may still be having issues scaling Gate-First to 20nm, and do not yet have a backup plan to transition to Gate-Last.
No, they planned to go gate-last at 20nm and that is probably the issue as they are new to it. They should have stuck to gate-first by the looks of things.
The stock has already found support at its 50 week moving average of 3.56 (see 2nd chart in intro box) so your 'realistic target' of $2.80-$3.30 may be too pessimistic unless earnings/guidance continue to disappoint. Of course there is a wall of dropping resistance above the stock too so expecting even a move over 4 maybe ambitious too. Ultimately the direction of travel all depends on whether the 50-200 DMA at 3.76-3.77 can be broken down, that is the crucial battle that lies ahead for this stock.
I think Goldmont's ipc will be greater (50%-55%) compared to Silvermont and its clockspeed lower than you think. It should though gain more clockspeed on the 10nm shrink. Basically IMO Goldmont should be a modern Core Solo (Yonah), the last of the triple-issue Pentium-M line.
Astrogenetix (update) – Astrogenetix is a biotechnology company formed to commercialize products processed in the unique environment of microgravity. Astrogenetix pursued an aggressive space access strategy to take advantage of the Space Shuttle program prior to its retirement in 2011. This strategy gave Astrogenetix unprecedented access to research in microgravity, as we flew experiments twelve times over a three year period.
In collaboration with NASA, NASA has engaged leading vaccine development experts through a premier educational institution to independently evaluate Astrogenetix’s platform with specific direction to aid in the filing of an Investigational New Drug (“IND”) application for Salmonella. While investment in Astrogenetix has been scaled back considerably until Astrogenetix’s platform is independently verified, the team is also evaluating a vaccine target for Methicillin-Resistant Staphylococcus Aureus (“MRSA”) based on discoveries made in microgravity. In December 2011, the Company negotiated a Space Act Agreement with NASA for a minimum of twenty-eight additional space flights.
http://www.sec.gov/Archives/edgar/data/1001907/000151597114000051/astc10q123113.htm
This quarter looks a little light too unless they are processing DMSP F19 and I am not sure of that. Work picks up in the next quarter (final quarter of FY2014 ending June reported September) when they process OCO 2 and WorldView 3. I also believe there is outstanding $3m+ worth of GSE revenue that could suddenly appear in any quarter.
http://www.spaceflightnow.com/tracking/index.html
TA wise as long as the stock can find support on a short-term moving average (e.g. the 75 MA is 2.38, 100 MA is 2.03) it is not a problem because half the numbers in those averages are under 1 so these averages will rise sharply back up to 3 over the next few weeks/months taking the pps with it.
You think K1 on 28nm will be able to offer the same perf/watt, or do you think it'll be at the same perf level but at a higher power envelope?
At least the latter if not the former.
Intel should have bought NVIDIA back when it was a $12/share stock.
Intel gets what it needs from its patent deal and no way was Huang going to be installed as Intel CEO so that pipe dream was never going to happen.
Should be fun to see how Cherry Trail does against its 28nm competition
I suspect only Nvidia's K1 will be in the same all round league.
They have delayed 20nm so they can make S5 on their imaginary 14FF Seriously who believed that story !
Digitimes (for once) nailed this 20nm Samsung yield problem right earlier.
The quarter was already known to be light in launch revenue and it did not help that the outstanding $3m+ GSE revenue is still outstanding. The gross profit was also only negative due to a non-cash stock compensation charge.
Stock back at 2.40
http://www.nasdaq.com/symbol/astc/after-hours
That's well over the 100 MA. This puppy is not going down without a fight.
Considering Intel has less volume/scale than TSMC and its 22nm is more complicated (FinFET) and denser (metal and especially gate) that's a pretty good showing and confirms what many of us have been saying for some time, Intel can do a 100 mm^2 22nm die for under $10.
It would be a good fit and Apple is looking for its own dedicated FAB.
When's the Kyro III coming out . Seriously, any plans to ever get back into discrete gpus ?
Intel insiders selling their shares is not a new occurrence, that's mainly why the buyback is there.