"TSMC is dedicated to providing foundry services for logic IC providers worldwide, and mergers in the industry will have little impact on the foundry's operations, said company chairman Morris Chang in response to media queries on Altera's recent announcement of being acquired by Intel.
Many market observers consider that Intel's acquisition of Altera, which makes use of TSMC's foundry services, may have an adverse impact on TSMC.
Intel could also rely on TSMC's foundry services in the future, Chang noted."
Over on digitimes.
Servers and the IoT. If that doesn't make sense to you then you don't have the brains god give a rubber duck.
The surprisingly smart part of the deal is how it gives Intel a decided edge over ARM's IoT capabilities. If Monk can't figure it out it's no surprise that the rest of the ARM pinheads can't either.
The idea that Intel will add a FPGA to a $2 microcontroller is laughable.
As for servers, there is a clear use case, but that's not going to keep the ARMy out. Why? The ARMy aren't focusing on compute, but IO,
Why not go for a macerating functionality and throw in some testing for blood while you are at it? All possible with a microcontroller as the logic is basic. Infrared sensor to detect someone coming, and weight sensors for the movement, then upload to your physician. You could always connect it to twitter and announce to the world the weight of your new produce. Or how about taking a photo and automatically sending to facebook to show all your friends? Nothing that a raspberry (no pun intended) pi couldn't handle.
Seriously, where is the need for an FPGA in any of it?
I can guarantee know more about IoT than you or some of the articles you have posted about. IoT isn't a new thing, it's a new marketing buzz term.
But you wont read this, as I'm on ignore...
I made a terrible mistake. Trying to dispel your IoT ignorance. Go away, earwig boy. If I wanted to play this stupid game I would just engage Twink.
Bless. You don't understand, it's OK, it happens.
The author is confused by several things (including the company that Intel actually bought), but let me give you a real world example. Type the following into your favorite search engine.
This is a $19 Wi-Fi enabled IOT development platform and contains an microcontroller and is typical of the kind of things that are being built and what the author is talking about.
The author is basically suggesting that the ARM microcontroller ( a 120Mhz ARM Cortex M3 with 128KB RAM) is replace by the (huge) Intel Quark and Intel add in a FPGA which "provides greater flexibility and faster time-to-market for customers". Lets ignore things like the die size (ie, cost), which markets could now be addressed with such a SoC that couldn't be addressed with the Photon above?
The examples, thermostats, toasters, LED lighting, subway systems (?), traffic lights, water hydrants would be addressed with something like the Photon. The only example that makes sense is "medical imaging systems" but here you aren't talking about a humble microcontroller...
[Once again, read the article. You can read, can't you?]
I have, but, like I said, can you give me an example. I can't think of one. You do understand what he was talking about, don't you?
BTW, the author has a terrible mistake on the first line. Did you spot it?
[Read the article, pinhead. Your ignorance of the IoT is only eclipsed by your unwillingness to change your ignorance. ]
I have, which is why I am asking the question. Why would you use FPGA in the IoT space? Can you give me an example?
Servers and the IoT.
Why would an embedded FPGA help in the IoT space?
[LOL. Seriously, you believe this will continue just like before? In that case I've got a watch, a bridge and some swampland you are going to be very interested in!!!]
Yes, I do actually. Do you know why? Intel said so. Why do you think they mentioned it explicitly in the press release...
Surprised ARM hasn't dropped below $50 on the Altera news.
Why? From Intel's own press release:
"Altera will become an Intel business unit to facilitate continuity of existing and new customer sales and support. Intel plans to continue support and development for Altera's ARM-based and power management product lines."
It is a 2.5 Ghz dual-issue out-of-order ARM chip i.e. one core will be slower than a 2.6 GHz Atom Avoton core. Even with 48 cores @100W it won't beat in throughput a similarly watt-powered Xeon set-up which will have much greater single-thread performance or 40 Avoton cores which you can also get for 100W. Thunder's only real purpose is to replace the similarly designed Cavium MIPS Octeon in-order chip in network chips.
I can't disagree with any of the above, but I think you are missing the target market that Cavium is going for with this type of chip. Basically, the workload wouldn't be CPU based, but IO based. Think something like AWS's S3 buckets, minimal CPU, but lots of IO overhead (basically, moving data from disc to network). Assuming these things have hardwired SSL acceleration board, they could do a nice job. I couldn't see the specs of the Asus motherboard, other than it being dual socket but I would expect it to have massive IO links.
There has to be a really, really compelling ROI reason not to use the industry standard x86 solution for enterprise use. Given the very efficient low-power multi-core products coming out of Intel I just don't see the necessary risk/reward benefit.
Why do you think Asus is doing this?
The first question many data center managers will ask is if it's VSphere compatible. If it's not it won't even be considered.
Not if you aren't a VMware shop:)
This isn't destined for typical corporate data centre use.
"TAIPEI, Taiwan, May 29, 2015 ASUS today announced new servers and server boards based on Cavium's ThunderX™ 64-bit, 48-core ARM® platform and designed to deliver extreme efficiency, expandability and enhanced-usability design for premium performance in data centers, high-performance computing and enterprise, as well as small and medium business (SMB) environments."
Ash speculates that the new mysterious Intel device might be the Galaxy Tab S2. And he suggested that the reason might be lack of capacity rather than the silliness that Monk suggested and presented as "fact"..
The "silliness" I was talking about is fact. Why would Samsung sell one device for $500 and another for $200 which contain the same SoC?
[Hahahaha. That's hilarious. I can't work out if you are being serious, or just trolling. ]
[Well, Samsung now has the incredible FinFET technology (according to you) and this incredible new smartphone (according to you) and well, incredible everything - which they didn't have in the past. And yet they are apparently doing Intel-based products. Of course, maybe you misrepresented all of this incredibleness and it is just the same as in the past. ]
I can't work out if you are being serious, or just trolling.
Why does Intel have a range of SKU's? Look at the number of Atom, Core and Xeon designs. Look at mediatek, look at QCOM. How many SKU's does Samsung have?
Samsung electronics (typically) produces one exynos version per generation, giving Samsung the device manufacturer little or no ability to differentiate it's offerings across the price/functionality curve. This is why you see Samsung using SoC's from a variety of firms.
Like I said, this is a good win for Intel, and it'll slot nicely into Samsung's mid range tablet offering. If Samsung starts to use a range of Intel SoCs across its product portfolio, then it would be very big indeed.