Haswell really brought out some experimentation by Intel, not seen since P4 end to Core transition; Paxville, Tulsa, Dunnington, Dempsey, Clarkdale, Bloomfield, Lynnfield.
Thus far for product that can be just about anything for everything; 4770, 4770K, 4570, 4670K, 4430, 4670 are the top supply grades. 4930 MX ranks 19th.
E3 becomes a sales outlet for the run's low performance sludge. Of all the design manufacturer's that invested in v3 board design, how many have their financial resources consumed in other developments like EP/EN v2, E5/E7 v2, Avoton and ARM Server?
Haswell-H MCP (a.k.a.) Crystal Well covered at Hot Chips appears arc'd before getting to the starting blocks;
At 348 mm^2 of silicon on interposer, one wonders if the L3 wasn't crippled and L4 doubled to accommodate, at these price points $383 to $657 which are competitive through economic profits for the top sellers, if Crystal Well might have been an applications performer? Estimated cost per unit is huge but a worthy experiment none the less.
i7/5/3 dual mobile SIP with integrated voltage regulator also shows design savvy consuming that competitive IP block. Over the production short run Intel charges $20 for voltage regulation over the standard quad's marginal cost to produce for nearly the same dice area; Dual w/voltage Regulation 181 v Quad at 177 mm^2. Unfortunately this part currently costs Intel $222 a pop. If any manufacturing operation can get the marginal cost of these mixed signal system's in package down its sure to be Intel.
So for workstation, performance and mainstream desktop, performance and main stream notebook and that illusive fourth generation ultra book, Ivy E Series saves the next three sales quarters along with its Xeon siblings.
Haswell full run on Intel experimentation comes in at Marginal Cost > Marginal Revenue by $10.15 each unit produced.
--IBM, Oracle and Fujitsu are increasingly using similar big iron techniques. The last three giants of scale up computing.
--It will be very interesting to see if Qcomm becomes a significant DSP player
--There's barely enough different microprocessor teams out that to call this a microprocessor event anymore. FPGA and comms and misc stuff has been part of the program for some time and its related and interesting, albiet the event has lost some of its intensity of a dozen competitors battling it out.
--It would be cool if thre ARM SoC folks would step up, especially as we get to 2014 and a 9-12 different server SoCs.
--There's a wonderful crew of regulars who you rarely see anywhere else. The networking is the best part.
Wow if we have such a high resolution camera that we can see twist of arms from distance and can move camera focus comfortably it would have so many applications especially when dealing with sports. There are so many times human empires or even the computer empire is not able to find fouls because of technology limitations so tis will be of so much need and help.
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.