brm, the key thing that I am interested in understanding is whether Intel can have a compelling competitive advantage by including 3DXP in the SoC.
Also, if Intel can reduce amount of DRAM by including eDRAM+3DXP, it looks like the market for DRAM will begin to contract. And more $ will funnel through to Intel for the eDRAM and 3DXP.
Which would be a great, big win for Intel.
I tend to agree with your direction of thinking, except I think MTBF will be longer than 2-3 years . We don't have the exact specs yet though.
Yeah, I agree. I think the useful life of NAND-based devices is well over that. Here is some anecdotal evidence of first generation iPhones (2007) still in use - http://www.marketwatch.com/story/some-apple-fans-stick-with-original-2007-iphone-2013-09-10.
So yes, it looks like there definitely will be a compelling market for SoC+eDRAM+3DXP in a single package.
Alex,I posted couple of days back that NVM could be on package or embedded with SoC with an excerpt from EE Times.
However, some folks have raised a concern that NVM has less endurance than DRAM, so putting NVM on to the package reduces the life of the device - which on face value seems to be a valid concern. Then again, we have NAND (which has less endurance than 3DXP) being soldered on to the board of mobile devices, so that concern appears to be invalid.
So, to me, it appears that for devices that have a life of 2-3 years, a SoC package with integrated eDRAM+NVM would work fine. When it fails, you just move over to a new device - with all your apps and data being on the cloud, it shouldn't be a big deal.
It might be a slightly different story for notebooks/desktops/servers where they may still have eDRAM on the CPU while having 3DXP (and NAND) on separate chips.
What do you think?
Alex, that for sure indicates that 3D-XPoint will be in the same package as the AP or on the same die. Having it as a separate chip will increase latency because 3D-XPoint is slower than DRAM, and as such will be a move in the wrong direction.
Will be great to see more info from Intel on this.
Better performance across the whole SoC range due to reduced latencies.
Intel can price lower-end SoCs more competitively since it will be capturing more of the BOM [AP+Minimal DRAM+Higher 3- XPoint (instead of NAND)].
For the higher-end SoCs, it can leverage the higher performance of more 3D-XPoint memory (instead of NAND) for higher prices. Again, can capture more of the BOM and leverage this advantage against ARM vendors.
I am pretty sure all this should already be in the works.
Further, it is critical to remember Intel can afford to price all these SoCs at very little margin (or even just above cost) to protect incursions into higher-end CPUs. That has enormous value from an investor point of view. This is what ARM vendors are up against.
[It would be absolutely phenomenal if Intel can come out with these designs over the next 6-12 months. Would be a death knell for ARM vendors.]
Intel did confirm that 3D XPoint manufacturing is compatible with back-end-of-line (BEOL) processing, which opens up the possibility of deploying 3D Xpoint memory on top of a plane of logic and as an embedded non-volatile memory option. But in email correspondence emphasized that the launch discussion is only about the technology as the basis of a discrete stand-alone memory.
Microsoft made quite a few announcements in the last couple of months, most of them related to Windows Phone and the company's strategy towards its smartphone business.
After the first announcements, many thought - and still think - that Windows Phone is dead and Microsoft will drop support for its own OS in the next few years.
Still, Microsoft has been pretty clear in the latest statements and said that the latest changes are meant to streamline its smartphone business rather than bury it.
That being said, Windows Phone fans should certainly be excited for what's to come since Microsoft confirmed it would focus on “killer devices” rather than tons of affordable smartphones.
The first flagship smartphones will be launched by Microsoft this fall and they will be powered by Qualcomm's Snapdragon chipsets.
However, rumor has it that the Redmond-based company has a few more surprises for Windows Phone fans starting early next year.
A Surface phone powered by an Intel SoC? Yes, please
According to WindowsCentral, Microsoft already has a smartphone in the works which will be equipped with an Intel chipset:
“We have already seen what Continuum for phones brings when Microsoft’s next flagships launch this fall. However, what about going further? Someone asked me about 'Intel-powered phones' and if I thought they were coming.
“I can now answer you: yes, early next year. Microsoft does have an Intel plan in the works. Would that make a good use of the Surface brand once you have an x86 phone that can be a true computer?”
Today, another piece of the puzzle has been revealed, which supports the above statement. It looks like Windows 10 Mobile already features support for Intel Atom x3 processor series, also known as SoFIA, WMPU reports.
A screenshot taken from the Microsoft.PhoneFM.xml configuration file in Windows 10 Mobile shows that Intel's SoFIA processor is definitely among the supported chipsets.
There is an article on a Intel SoFIA-based Windows Surface phone coming out early 2016 for around $80. I will try to find the article and post it.
Unfortunately, MS seems to be making the same mistake with the Qualcomm/Continuum project that they did with RT. Microsoft is indeed a slow learner!
Asus Zenfone 2E now available via AT&T GoPhone for $119.99
Asus today announced that it is releasing a budget-focused device to AT&T’s prepaid GoPhone service today dubbed the Zenfone 2E. The device is very similar to the Zenfone 2, but less powerful and aimed at the budget-conscious customer.
Design-wise, the Zenfone 2E takes many cues from its more expensive sibling, but features a slightly smaller 5-inch display at 720p resolution. The Zenfone 2 features a 5.5-inch display at 1080p. The Zenfone 2E also features an 8MP rear facing camera and a 2MP front-facing shooter. The device is powered by a dual-core Atom processor paired with 1GB of RAM. The device ships with 8GB of onboard storage, but thankfully features a microSD card slot for further expansion.
The Zenfone 2E is available today in Walmart and Target retail stores and will come to AT&T locations on July 24th. The device will run you $119.99 and includes two interchangeable rear backing cases in white and black.
We’ve got a full review coming right around the corner.
Good post mega.hurts!
brm, in case you missed it, this sentence directly addresses your comment: "It may be fair to question why other vendors don't have phones based on the same parts (yet), but a number of those possibilities have nothing to do with the SoC's performance. The benchmarks are well documented."
AnandTech smacked you brm. At the very least, you should cease posting similar comments in future or you risk coming across as someone posting glibly without regard to the facts.
TSMC cuts 2015 chip growth outlook
Taiwan Semiconductor Manufacturing Company (TSMC), at its July 16 investors meeting, gave a less-optimistic outlook for the third quarter of 2015, and has revised downward its forecast for 2015 global semiconductor and foundry growth.
TSMC co-CEO Mark Liu indicated the company now forecasts the global semiconductor market will grow 3% in sales in 2015, and the foundry segment will see a 6% increase. TSMC previously estimated sales in the global foundry market would rise 10% in 2015, and the overall chip sector 4%.
Nevertheless, TSMC reiterated its target of double-digit growth in 2015 sales, Liu said.
"The end market recovery is not as strong as was expected earlier; customers continue to remain cautious in inventory management. Combining these factors with customers' product transition, demand for TSMC wafers in the third quarter is expected to recover only modestly," said company CFO Lora Ho.
TSMC forecast its consolidated revenues for the third quarter of 2015 will be between NT$207 billion (US$6.65 billion) and NT$210 billion, representing growth of up to 2.2% on quarter.
Global demand for smartphones has decelerated significantly. Inventory levels particularly in China were already high in the first quarter of 2015, but have reduced 50%, TSMC noted. However, excess inventory at handset companies might not be able to ease until the fourth quarter of the year, TSMC said.
TSMC saw its second-quarter revenues decrease 7.5% sequentially but meet its guidance. In the second quarter, TSMC's business was negatively impacted by customers' cautious inventory management and a less favorable exchange rate, Ho noted.
Sales generated from the communication sector accounted for as high as 62% of TSMC's total wafer revenues in the second quarter of 2015.
Looks like the Smartphone and Tablet markets is getting to be challenging due to saturation and macroeconomic conditions. Given that the prices and profitability of the SoC market is relatively low compared to the PC market, I'm pretty sure that manufacturers who end up with loads of inventory will have a difficult time climbing out of this mess...it is getting interesting to say the least.
Note: TSMC has been cutting its 2015 forecast and is quoted in this article as saying that "....global smartphone demand has decelerated significantly...".
Is BK stepping aside to watch this market implode?]
MediaTek likely to cut shipment target for 2015
MediaTek is likely to revise downward its handset-IC shipment target for 2015 to about 400 million units from 450 million, a 10% cut, according to market sources.
After experiencing a particularly weak first half of 2015, MediaTek will see orders rebound in the third quarter thanks to a pick-up in smartphone demand from China and emerging markets, as well as brand vendors' roll-outs of new products, said the sources. However, unfavorable macroeconomic conditions will cast a shadow on the demand outlook for the fourth quarter.
MediaTek is expected to see orders rise about 20% sequentially in the third quarter, but orders placed thus far for the fourth quarter have turned weak, the sources indicated.
MediaTek president Hsieh Ching-chiang said previously that customer demand started picking up in May and has strengthened since June, and expressed optimism about the company's performance in the third quarter.
Foundry chipmaker TSMC suggested at its recent investors meeting that global smartphone demand has decelerated significantly. Handset companies are struggling with excessive inventories which are unlikely to reduce to an adequate level until the fourth quarter, TSMC said.
Contd. in Part 2
brm: We (and I am including myself) constantly make the mistake of extrapolating our own preference to everyone else.
You admitted in your post above: "I can buy the argument that there is a sizeable chunk of the market who would like to have full fat windows running on a phone.".This is Chunk 1.
And then there are folks who insist that there is a huge chunk of users who can do very well with just a smartphone or tablet - and that the extra power and performance of a PC is not really needed. This is Chunk 2.
I was just combining these 2 chunks to say that these folks can really benefit from "OCCASIONAL" use of the smartphone as a PC with docking devices or a couple of peripherals like displays and keyboards. That would be handy for both Chunk 1 and Chunk 2.
One other form factor that I have been mulling about is as smartphone dockable into a laptop. This is along the lines of the Asus Padfone, but the phone docks into a laptop instead of a laptop. Perhaps in future, we will see some SoCs that can throttle up and down to accommodate both usages.
Please note that I'm not including medium-to-high end PC use case. For these folks, the PC form factor is king of the hill as you mentioned.
Actually, I think you unwittingly admitted to the versatility of a laptop :)
However, using a laptop as a phone is not a viable idea. So I do see a future where the peripheral devices can be provided in different environments and folks with Windows phones can just hook into them as and when necessary. Offices, airports, etc. can provide working kiosks with a display and keyboard. Hotel rooms can just add a keyboard in every room with the TV acting as a display (or they can just add another display on the desk).
I think the future is going to be very different from what we can envision...
As long as anyone needs Office, Microsoft will have some leverage over the customer.
Having said that, the leverage may not be as much as "mega.hurts" thinks. But it may not be close to zero as you seem to imply.
There may be a fairly decent size segment for whom Windows Phone on x86 can very nicely complement their laptops. How Microsoft is going to make a compelling case for this segment remains to be seen.
RISC vs CISC: What's the Difference?
Analysis of ARM, X86, MIPS designs shows no difference
A new study comparing the Intel X86, the ARM and MIPS CPUs finds that microarchitecture is more important than instruction set architecture, RISC or CISC.
If you are one of the few hardware or software developers out there who still think that instruction set architectures, reduced (RISC) or complex (CISC), have any significant effect on the power, energy or performance of your processor-based designs, forget it.
Ain't true. What is more important is the processor microarchitecture — the way those instructions are hardwired into the processor and what has been added to help them achieve a specific goal.
"Based on this study, developers can safely consider ARM, MIPS, and x86 processors simply as engineering design points optimized for different levels of performance," he said. “There's nothing fundamentally more energy-efficient in one ISA class versus another."
In the concluding paragraph of the report, the authors write: "It appears that decades of hardware and compiler research has enabled efficient handling of both RISC and CISC ISAs, and both are equally positioned for the coming years of energy-constrained innovation."
Sources: EE Times
You're right., my bad...should have had my coffee before responding. Intel is doing a lot of mixing and matching GPUs, should be easier to swap it out if Intel thinks it is worthwhile to do so.
You're right....with SoFIA, not ARM but hurting all lower-end ARM SoC vendors.
Over time if and when x86 core is substituted for ARM, will be hurting ARM as well. Don't mind Intel handing a few pennies over to ARM if it is not worthwhile to add x86 in place of ARM core :)
SEJ & Alex, appreciate your thoughts. Regarding Micron - Flat bits, lower ASPs, staggered quarter vis-a-vis Intel - too many moving parts to accurately assess Intel Q2 results. Guess we have to wait till Intel announces in a couple of weeks.