Slashdot Mirror

http://www.eetimes.com/author.asp?section_id=36&doc_id=1323406&_mc=RSS_EET_EDT

Put URL between "<URL:" and ">" to make it a link: "<URL:http://www.eetimes.com/author.asp?section_id=36&doc_id=1323406&_mc=RSS_EET_EDT>". Please don't be so lazy.

That's true for active power. (V^2/R). For leakage power, it's even worse. That looks closer to exponential. I've seen chip for which leakage accounted for close to half the power budget.

Supposedly FinFET /Tri-gate will help dramatically with leakage. We'll see.

http://www.eetimes.com/document.asp?doc_id=1323507
http://tech.slashdot.org/story/14/08/14/1921238/can-our-computers-continue-to-get-smaller-and-more-powerful

The Limits of Moore's Law Limits

TI+ARM is alive and well.

More like on life support since 2012: http://www.eetimes.com/documen...

The AM437x is nearly 18months late, and likely was far along in the pipeline when those layoffs happened. TI's OMAP roadmap has been slipping for the past two years, with many processors there will never get released.

http://www.eetimes.com/documen...

We're already at the point where 22nm components are more expensive per transistor than those at 28nm.

Previous shrinks lowered the cost of each transistor. It doesn't look like it's going to happen after 28nm.

What, no GPS transmitter in the filament of each paper Euro? Amateurs.

They have planned to add RFID. However AFAIK this has never been realized (yet).

Fluke owns Wavetek

http://www.eetimes.com/documen...

Those were decisions that should have made the Xbox One cheaper. It's basically the same architecture as the original Xbox 360, and is well-understood; by comparison the PS4's GDDR5 is luxuriously expensive. Kinect is definitely to blame.

While GDDR5 is definitely more expensive, the price difference isn't that massive, at least when you are a company as large as Sony or Microsoft with the corresponding bulk purchasing power. This estimate indicates that Sony's 8GB of GDDR5 costs about $62, compared to $39 for Microsoft's 8GB of DDR3. Add to that the fact that Microsoft is paying more for a larger APU die to offset the RAM's weakness: roughly $132 compared to $121. (Those figures are estimates, but we know that the XB1 APU die is 363 mm^2, compared to 348 mm^2 for PS4.)

So when you factor the larger and more expensive die into the equation, Microsoft saved a grand total of $12 a unit by going with DDR3 – and in the process, reduced their graphics performance significantly. Like I said, the only sensible explanation is that the Microsoft designers drastically overestimated the cost savings of skimping on primary system RAM, and probably also underestimated the performance hit it would cause because of the die space trade-off.

EETimes has a more useful article. This is more like a reversible fuel cell. The working fluid is pumped through the cell, where a chemical reaction occurs. The process is reversible. So there's a "charged" fuel tank, a "discharged" fuel tank, pumps, and plumbing. No info yet on the energy density of the "charged" fuel tank, which is the big question.

I read the article. The article is wrong. I was around when they started suing, as your UID shows you should have been. Here is a contemporary article in EE Times about the first lawsuit in 2000, against Hitachi. The article's wording makes it clear that this is the first lawsuit. This was big news on Slashdot in 2000, where an article from 2000 whose wording (and comments) also make it obvious that the lawsuits had just begun. Here is a slightly earlier Slashdot article about Samsng and Rambus whose comments also make it obvious that no lawsuits had begun yet.

Why is it that Ethernet can push 10gbit/s full duplex over a 100m unshielded cable,...

Precision wire twisting eliminates crosstalk and unequal induction.

Seriously. It's all in the twist.

...but USB (2) can only push 480mbit/s over a 5m cable (or 4gbit/s over the same length with USB 3)? Why the hell didn't the USB designers take a cue from the Ethernet cable designers?

Probably because USB is an evolution of serial communication between two endpoints (think RS232c and friends, or the ancient and beloved 20 mA loop) which isn't remotely the same paradigm as CSMACD networking.

Is this true, or are there more important factors to consider when choosing an SSD brand/model?

Is this true, or are there more important factors to consider when choosing an SSD brand/model?

It wasn't a sensor bit error that it failed to guard against. The control values I referred to are those in RAM, used by the software. The RAM apparently wasn't parity protected, and a bit-flip in the right word could cause uncontrolled acceleration. It wasn't the only thing that could cause havoc; there were race conditions and stack overflows in the code, apparently, and those were more likely the sources of actual, observed UA.

This lengthy article at EE Times digs into some of the details. The main quote, though, is on page 3:

Memory corruption as little as one bit flip can cause a task to die. This can happen by hardware single-event upsets -- i.e., bit flip -- or via one of the many software bugs, such as buffer overflows and race conditions, we identified in the code.

There are tens of millions of combinations of untested task death, any of which could happen in any possible vehicle/software state. Too many to test them all. But vehicle tests we have done in 2005 and 2008 Camrys show that even just the death of Task X by itself can cause loss of throttle control by the driver -- even as combustion continues to power the engine. In a nutshell, the fail safes Toyota did install have gaps in them and are inadequate to detect all of the ways UA can occur via software.

I don't think that article pointed out this other detail:

Although the investigation focused almost entirely on software, there is at least one HW factor: Toyota claimed the 2005 Camry's main CPU had error detecting and correcting (EDAC) RAM. It didn't. EDAC, or at least parity RAM, is relatively easy and low-cost insurance for safety-critical systems.

This particular set of problems at Toyota was very interesting to us at work. I'm just now starting to work with our safety critical team that sells hardened controllers into the automotive market. They include all sorts of hardware failsafes, including ECC, lockstep execution between parallel cores, etc.

Yes, that is how encryption works. But if your key is large enough, the time & energy to brute force it will take much longer than your lifespan. As an example I just googled, brute-forcing AES-128 at 10 Petaflops would take 10 quintillion years (10^18). http://www.eetimes.com/document.asp?doc_id=1279619

The _real_ concern is that the NSA knows of weaknesses in these encryption schemes, and doesn't have to brute force it.

It depends on the fuel cell and the fuel. Platinum is used in lightweight, low-temperature fuel cells meant for rapid load changes (as in cars and such). If you are running from natural gas or other fuel sources, and are running with a high-temperature fuel cell, you don't need platinum.

But people are forgetting about the distribution costs with electricity. A significant portion of the energy is lost as heat in the distribution system - both as I2R losses in the wires and inefficiencies in the transformers. Unless you have a big leak in the piping, shipping natural gas around is basically loss-less aside from pumping costs.

Fuel cells running more or less at steady state at the point of use are very economical.

http://www1.eere.energy.gov/hydrogenandfuelcells/pdfs/fccs_omaha10.pdf

http://www.eetimes.com/document.asp?doc_id=1139680

http://www.usatoday.com/tech/news/story/2012-06-21/ebay-fuel-cells/55732562/1

to users of the almost extinct XP.

If by "almost extinct" you mean "Still installed on over half the computer in Asia," then sure. Bonus: It's still installed on over a third of machines worldwide I'm sure China will appreciate yet another American trying to meddle in their developing economy by giving them an operating system that no other major country would be expected to use as a replacement desktop, simply because we were nice enough to provide it free of charge. There's no possible way this do-good attempt to "save" the good people of another country... because as we all know, Western culture never tries to force other cultures to do what they want "for their own good".

Get a little international perspective, Slashdot. It's not like Microsoft practically gave away XP to developing countries to lock them in, and then jacked the price up for the upgrade, while giggling maniacally like some evil overlord. Jeez... why does the rest of the world hate imperialism so much? It's not like it ever enviously eyed their booming economy and thought; What can we do to cash in on this without looking like total dicks? Oh! I know... how about we pull support for their dominant operating system and force them to spend hundreds of millions on upgrade fees!

Snark aside, this disproportionately targets developing economies and non-western countries. Anywhere else, this would incite comments about racism, cultural warfare, etc., but since it's just an innocent tech giant all I've heard is crickets. Put yourself in the shoes of the rest of the world though; They have aging computers that can't run Windows 7 because of the significantly higher hardware requirements, and while they produce most of that hardware, they can't buy it because their workers all make pennies on the dollar. Are these people just going to drop off the internet, crawl down a dark hole, and die as a service to the rest of us, who hate the poor in the third world? Probably not: Instead, they'll form the backbone of supermassive new botnets, without any way to secure those systems, it'll become a cesspool of all manner of digital evil.

But at least Microsoft's profit margins will be up, so there is that.

"Engineers are now starting to get paid for their true value, which arguably has not been case for a long time, but it is now, and Google is at heart of this. Google discovered an algorithm change can generate another $100 million in revenue. So now companies are more willing to have superstars, and there are engineers at Goggle making tens of millions of dollars."

http://www.eetimes.com/document.asp?doc_id=1319417

Please see http://www.eetimes.com/document.asp?doc_id=1280095 for more information on interesting developments with regards to automotive usages of Audio Video Bridging in Infotainment and control - http://avb.statusbar.com/faq.html - typically using Linux.

A lot of companies that used to maintain their own fabs have closed them over the years. Trying to keep up with the leaders in process technology is very expensive. It has been a long time since it was even as cheap as $1,000,000,000. Not many companies can afford to build one.

Semi industry fab costs limit industry growth

By 2020, current cost trends will lead to an average cost of between $15 billion and $20 billion for a leading-edge fab, according to the report. By 2016, the minimum capital expenditure budget needed to justify the building of a new fab will range from $8 billion to $10 billion for logic, $3.5 billion to $4.5 billion for DRAM and $6 billion to $7 billion for NAND flash, according to the report.

It used to be that companies could leverage their own fabs for competitive advantage in process or design technology, or simple scheduling. Not any more. Now you outsource the fab to one of the big providers and get in line. More and more of the fabs are outside the US.

Some of the smaller old fabs get retargeted to specialty products, but even that tends to die eventually.

That's a bit disingenuous. You can't pick out a single stage of the process and compare that while ignoring everything else if you want an honest assessment of the efficiency.

Let's make up an example. Let's say you have a source of fuel, a power plant that can burn that fuel, a testing ground that's 100 miles further away from the fuel source than the power plant, and two vehicles that can utilize the fuel. Internal combustion cars are generally about 15% efficient. Electric engine cars are 85-90% efficient. Fossil fuel power plants are about 33% efficient. Your transmission numbers are 99.86% for gasoline and 99.25% for electricity over 100 miles.

So overall efficiencies are:
Gas: 1 * 0.9986 * 0.15 = 0.14979
Electrical = 1 * 0.33 * 0.9925 * 0.88 = .28822

So over a distance of 100 miles Electric cars are still almost twice as efficient, even with the extra losses in transmission. (Admittedly this is for "normal" internal combustion cars, i don't have the figures to hand for the average efficiency of hybrid cars.)

Doing a little quick math (it's been forever since i've had to solve for a variable, so i'm just plugging it into a spreadsheet) it looks like the break-even point is about 10,700 miles. So if the distance from the fuel source was over 10,700 miles, you'd be better shipping the fuel to the car rather than converting it to electricity on-site and transmitting it to the destination. Though obviously over such an extreme distance a lot of other factors would come into play and overwhelm the simple equation.

Sources:
http://consumerenergycenter.org/transportation/consumer_tips/vehicle_energy_losses.html
https://en.wikipedia.org/wiki/Fossil-fuel_power_station
http://www.eetimes.com/document.asp?doc_id=1273932

There may be more accurate numbers out there, so the exact outcome might differ, it's clear that better efficiency in just a single stage of the operation does not dictate an overall higher efficiency.

In fairness to AnTuTu they released a new version which tries to rectify the problem:
http://www.eetimes.com/author.asp?section_id=36&doc_id=1318894&

You will never be able to get 1Gbps over a 100m long CAT5 cable

Yes you will. The 1000BASE-T (802.3ab) specifications specifically say 100m on CAT-5 cable, and plenty of people have done so, successfully.

http://www.eetimes.com/document.asp?doc_id=1205868

You might have range issues if an idiot patched your CAT-5 cables and did too much untwisting, or made similar mistakes. But if a professional installed it, they were tested and certified (by a certifier that cost several thousand dollars) to fully meet CAT-5 specifications, before anyone signed-off and paid for the installation.

Now, what you MIGHT have been thinking of is 10GBASE-T, which can only do 55m over CAT-6, an needs CAT-6a for 100m, but that's an order of magnitude difference, there.

You made me curious so I googled it: TechInsights' teardown uncovered within Kinect a Marvell PXA 168 applications processor.

This isn't a new issue to people in the industry. Here's a more useful article from last year: "Is the cost reduction associated with IC scaling over?" "Clearly, dimensional scaling is no longer associated with lower average cost per transistor."

The cost of wafer fabs has been going up with each generation. Intel says that a cutting-edge fab now costs upwards of $10 billion, twice the previous generation. That's why higher densities no longer reduce cost. The upper limits of optical lithography are being reached because light, even "deep ultraviolet" light, is too coarse a tool. "Extreme ultraviolet" (soft X-rays, really) are being tried to get down to 10nm or so, but the processes are currently slow and barely work. Electron beam machines, which can go below 10nm, have been around since the 1980s, but they work by writing the chip with an electron beam, not with a mask, which is very slow for a production process.

This is for mostly-static memory. For active transistors, as in CPUs, heat dissipation is already limiting density. CPU clock speed maxed out between 3 and 4 GHz several years ago. (Yes, 8GHz has been achieved with an AMD CPU running in liquid helium. So?)

With the upper limits of speed and density in sight, work is now focusing on reducing cost and power consumption. Hence the push to use ARM CPUs in more applications.

What evidence do you have that more people are applying for entry to engineering programs? It seems to me that your premise is likely only true if universities could be shown to enrolling significantly more engineering students over the years. If the number of people getting engineering degrees is any indication, the information in this article would indicate that universities have not significantly increased their engineering school enrollments. If anything they may be enrolling fewer engineering students overall.

Wow, you sound like a real industry expert, someone I'd certainly trust to evaluate new technological innovations.
No one could possibly ever need microservers for anything.

I did a load of background reading on this yesterday so here's some interesting related material. One interesting source is the NASA guidelines for li-ion use in space :
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20090023862_2009023573.pdf

Now NASA I think have a pretty good track record of thinking technologies through carefully... (By the by, did you know that GS Yuasa also have a contract to supply their li-ion batteries to NASA for use in the ISS?)

Also, did you know that prior to the 787 the Cessna CJ4 was the first civili aircraft to utilise li-ion batteries (supplied by a123). In 2011 there was a fire onboard one whilst it was connected to a ground power unit. As a result the FAA ordered all 42 in operation to be changed to conventional ni-cd or lead acid.
http://www.gpo.gov/fdsys/pkg/FR-2011-11-01/html/2011-27596.htm

This is interesting as it's similar i.e. on the ground. This of course *could* be coincidental.

Next up are lots of interesting pictures from the NTSB investigation. Much as I HATE to link to the Daily Mail (normally a pretty retarded publication) I couldn't find any other pic sources. Bizarre :
http://www.dailymail.co.uk/news/article-2268152/Boeing-787-Dreamliners-burnt-battery-spewed-molten-electrolytes-reveal-investigators.html?ito=feeds-newsxml

And some great source material from the NTSB themselves :
http://www.ntsb.gov/investigations/2013/boeing_787/boeing_787.html

And the NTSB update on the investigation (including some samples of their cell CT scans) :
http://www.ntsb.gov/investigations/2013/boeing_787/JAL_B-787_1-24-13.pdf

NTSB Primer on li-ion battery tech :
http://www.ntsb.gov/investigations/2013/boeing_787/Primer_LIB_Technology.pdf

One of the theories being talked about are the fact that the li-ion batteries that Boieng (via Thales) decided on are based on a lithium cobalt oxide cathode which is old tech and regarded as not exactly the safest variant of li-ion technology out there :
http://www.designnews.com/document.asp?doc_id=257987
and via a translation :
http://translate.google.com/translate?sl=no&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&eotf=1&u=http%3A%2F%2Fwww.tu.no%2Findustri%2F2013%2F01%2F17%2Fher-er-dreamliner-problemet

This EEtimes article has some interesting comments :
http://cdn.eetimes.com/electronics-news/4405441/787-Dreamliner-investigation-probes-battery-charging-electronic

And some info from GS Yuasa :
http://www.s399157097.onlinehome.us/SpecSheets/LVP10-65.pdf

All interesting stuff. Personally I think they shouldn't have been allowed to 'trial' li-ion on such a big aircraft especially after the cessna incident. Trying so many new tricks at once isn't wise - as engineers always say, just change one thing at a time...

Well, at a minimum, MSFT is in that group too.

http://www.eetimes.com/electronics-news/4204863/Microsoft-takes-ARM-license

Oooh, even better... get him the plans and parts for a nixie-tube clock!

http://www.eetimes.com/electronics-news/4396235/Microchip-s-Nixie-Tubes-light-up-Design-East

The article is remarkably lacking in technical details.

This article from two years ago is a little more detailed: http://www.eetimes.com/electronics-news/4211151/IBM-debuts-CMOS-silicon-nanophotonics
or this press release: http://www-03.ibm.com/press/us/en/pressrelease/33115.wss

You can't carry atomic clocks in your pocket, they're a tad too big for that.

Actually, it seems you can. And even older generations weren't that huge - certainly too big to carry in your pocket, but not impossible to mount in a vehicle, aircraft or ship. E.g. this HP 5071A which can be yours for the bargain basement price of just $27.500.

So maybe some GPS receivers can indeed get a 3D fix with 3 satellites, when combined with (or including) an atomic clock.

You have a point, as time would be a parameter of the simulation, but to the extend that we can make any statements about such an "outer" computer it'll still follow the same rules as our own Turing machines, everywhere QM is simulated in the "virtual" universe it'll quickly run up exponentially growing resource consumption. And why exactly do we assume an "outer" classic computing machine rather than a Quantum computer or a good old analog computer?

Although the prevalent digital computers of our time are all Turing machine equivalent, there is no such thing as just one computing paradigm.

This entire enterprise reminds me of the obsolete yestercentury assumption that the universe was functioning just like the innards of a mechanical clock.

Intel's fab advantage does not necesarily mean they win the mobile space. In fact, the fabless model is better in some ways, in terms of IP availability for SoC integration.

Here's a very interesting series of articles: http://www.eetimes.com/discussion/other/4235936/Viewpoint--How-Will-The-Chip-Wars-Be-Won---Part-1 and http://www.eetimes.com/discussion/other/4236454/-How-Will-The-Chip-Wars-Be-Won----Part-2

Intel's fab advantage does not necesarily mean they win the mobile space. In fact, the fabless model is better in some ways, in terms of IP availability for SoC integration.

Here's a very interesting series of articles: http://www.eetimes.com/discussion/other/4235936/Viewpoint--How-Will-The-Chip-Wars-Be-Won---Part-1 and http://www.eetimes.com/discussion/other/4236454/-How-Will-The-Chip-Wars-Be-Won----Part-2

Cold fusion experimentally confirmed

U.S. Navy researchers claimed to have experimentally confirmed cold fusion in a presentation at the American Chemical Society's annual meeting.
"We have compelling evidence that fusion reactions are occurring" at room temperature, said Pamela Mosier-Boss, a scientist with the Space and Naval Warfare Systems Center (San Diego).

Article from 2009.

Oh, and one more thing I forgot to mention, that's just funny, and probably has nothing to do with anything. Both his mother and father were identical twins, and their twins married each other, making two (nearly) identical couples genetically. There's even more, but it starts to get too personal. Ken's story blows me away. I assume many readers will have come to the very reasonable conclusion that either I'm pulling your leg or am dillusional, but this is a case of reality cooler than fiction. Here's a link about Synplicity, and here's one about Ken.

Now days, SFAIK, he's devoting his brain cycles to physics as a Ph. D. student at UC Berkeley. I keep waiting to hear about the mysteries of the universe he's decoded. One more thing about Ken: he rarely would read papers in our field. He would say that reading other people's solutions to hard problems would bias is own thinking, causing him to get trapped in the same solution space as everyone else. I read a few of his best algorithms, and I have to say he seemed to be able to consistently out-invent the combined rest of the world.

"I think the jury did an admirable job making sense of the case they were given. They certainly did better than much of the tech media, which have made a complete mess of the verdict."

The jury *completely* screwed up. They started by ignoring the prior-art argument[1] samsung made, and then the foreman proceeded to sway everyone[2] with an "I got my own patent so listen to me" bullshit. The jury was an ill-fated catastrophe from the beginning of deliberations.

[1] http://www.techdirt.com/articles/20120826/23534320161/applesamsung-jurors-admit-they-finished-quickly-ignoring-prior-art-other-key-factors.shtml

[2] http://www.eetimes.com/electronics-news/4394863/Jury-foreman-recounts-Apple-vs--Samsung-case

Here is an EE Times article on the same subject. The article and commenters agree people using mass transit (mostly outside the US) want to minimize carried weight and prefer a single device with a large screen instead of two devices (pad and phone). If you are already carrying a day pack, the size is less of a problem than weight from two batteries.

my samsung galaxy s II is plenty big and im happy as a clam with it and how BIG the screen actually is maybe you're iphone cant hold up or you're blackberry isnt cool but guess what androids with big screens is cool with us.

Here is an EE Times article on the same subject. The article and commenters agree people using mass transit (mostly outside the US) want to minimize carried weight and prefer a single device with a large screen instead of two devices (pad and phone). If you are already carrying a day pack, the size is less of a problem than weight from two batteries.

Don't worry, ASML is offering equity stake to TSMC and Samsung as well: http://www.eetimes.com/electronics-news/4389932/ASML-in-talks-with-Samsung-TSMC-on-equity

Freescale is working on such a chip: http://www.eetimes.com/electronics-news/4375606/Freescale-adopts-ARM-cores-in-QorIQ-line

int i = 0;
int f(){i++; if(i == 1) return 5; else return 6;}

int g() {i++; return 6;}

float h(int i, int j){return 1/(i-j);}

int main()
{
h(f(),g());
return 0;
}

The code in main() looks innocent at first -- but it has undefined behavior. This is something that is easy to accidentally write, too; you could call any number of standard library functions that update some global state. The C standard, section 6.5.2.2, says the following about this sort of thing:

The order of evaluation of the function designator, the actual arguments, and subexpressions within the actual arguments is unspecified, but there is a sequence point before the actual call.

Things are even worse in C++; you might not even realize that you have function calls in your argument list in C++ e.g. you might be calling a copy constructor for a smart pointer.

Of course I should actually link to the random article. http://savolainen.wordpress.com/2011/09/ This article also has some more information, but less cool diagrams. http://eetimes.com/electronics-news/4376121/Applied-tips-dielectric-etch-tool-for-3-D-NAND-production- Disclaimer: I work for Applied Materials (not for the etch division though).

It's about diversity, no redundancy.
For semiconductor manufacturing, for example. There's quite a few plants around the world, not that much of small process nodes. If a major earthquake hits Taiwan, shutting down TSMC and UMC factories, you'll notice the effect all around the world pretty fast.
This http://eetimes.com/ContentEETimes/Images/120514_icInsights_micron_800.png should give an idea.

Samsung/Hynix and HP have stated they they are producing the machinery for mass-production of ReRAM and expect it to be in retail by 2013 for storage and 2014 for system memory if all goes well.

Here's one quick google that I found. I read some other one, but yeah..
http://www.eetimes.com/electronics-news/4229171/HP-Hynix-to-launch-memristor-memory-2013
10/6/2011 HP, Hynix plan to launch memristor memory in 2013

I'm surprised that *no-one* has mentioned ExpressLogic vs Green Hills (see http://www.eetimes.com/electronics-news/4061092/Express-Logic-seeks-injunction-against-Green-Hills) at all, during this sordid saga.

It seems that was probably one of the earliest cases regarding the legitimacy of API cloning (GH tried to implement a "ThreadX"-compatible API on top of one of their proprietary RTOSes), and eventually lead to Green Hills winning against ExpressLogic (see http://www.zdnet.com/blog/gardner/ruling-expressly-denies-express-logic-its-copyrighted-api-logic/2530).

That said, I also thought of the Sony vs Connectix case - but there, Connectix used reverse-engineered PlayStation BIOS code heavily in their own product, if I remember correctly.

The 64 bit ARM architecture for server CPUs is much more interesting ...