Analyst slams Isuppli for Moore's Law death prediction
AN ISUPPLI REPORT predicting the death of Moore's Law by 2014 has analysts close to Intel breaking out in hives.
"Don't believe the idiots writing these things," we were told by an industry insider after asking him whether he thought Isuppli's view that high costs of semiconductor manufacturing equipment would kill off Moore's Law in just five years was accurate.
In a statement, Isuppli's director and chief analyst, Len Jelinek yesterday suggested that "The usable limit for semiconductor process technology will be reached when chip process geometries shrink to be smaller than 20 nanometers (nm), to 18nm nodes" and that "at those nodes, the industry will start getting to the point where semiconductor manufacturing tools are too expensive to depreciate with volume production, [that is], their costs will be so high, that the value of their lifetime productivity can never justify it."
But according to In-Stat's chief technology strategist, Jim McGregor, a lot of companies are facing financial difficulty at the moment due to over expansion in the industry and the general economy, and this "has nothing to do with this argument."
McGregor told the INQ, "even Gordon Moore has predicted the end of Moore's Law twice, but he was wrong both times." Reassuring indeed.
According to McGregor, Moore's Law - the observation that the number of transistors that can be placed on an integrated circuit doubles roughly every two years - is "a key fundamental driver of the industry from an economics perspective." Future challenges, says McGregor emphatically, will most likely be down to technology issues rather than equipment costs, so Isuppli is spouting nonsense.
Indeed, every new technology goes over some road-bumps, especially involving start-up costs, but these tend to drop rapidly once moved into regular production. "EUV [extreme ultraviolet] will likely be the next significant technology to go through this cycle," McGregor told us.
McGregor did concede that the lifecycle of certain technologies is being extended by firms who are in some cases choosing not to migrate to every new process node, but he maintained new process tech is still the key driver of small design geometries, including memory density, logic density, power consumption, etc.
"Moore's Law also improves the cost per device and per wafer," added McGregor, who also noted that "the industry has and will continue to go through changes because of some of the cost issues." These include the formation of process development alliances, like IBM's alliances, the transition to foundry manufacturing, and design for manufacturing techniques like computational lithography.
"Many people have predicted the end of Moore's Law and they have all been wrong," sighed McGregor. The same apparently goes for those foolhardy enough to attempt to predict changes in the dynamics of the semiconductor industry.
"There have always been challenges to the semiconductor technology roadmap, but for every obstacle, the industry has developed a solution and that will continue as long as we are talking about the hundreds of billion of dollars in revenue that are generated every year," he concluded. µ
Share this:
ShareBusiness Intelligence Software
Customer Relationship Management (CRM)
Enterprise Accounting Software
Enterprise Resource Planning (ERP)
Enterprise Systems Management
From Control Data transistors 1959 to MainFrame ibm 360 developement, look at number 3.(today 6) Yet, it took little time to go to memory that single trace scope had built in 1947, recovered while in service. Question:how to slow these Amphetamized people down, as youth had NO chnace to even live. Wala: Moores Law. Start Big Stay big, for long time. Yet true science, mainframe quickly turned to solid state memory & then matrix thru ssd cpu integration. trouble. Most of needed hardware to matrix wasn't invented & whole smear was TOP SECRET. Pentagon/Alliant TS. Input: Whittier, That Standard Bearer That Changes with Every Defeat.
theINQ Rev.13.64.08.000x NEXT: Heck, Hexadecimal Code Was Just Being Sorted into being. Ms/ Hooper trying desparted To Knave Something Workable to Industry. As Portable, Standardized Software? In name Fortran. You Pay For What You Don't UnderStand. 'Ole CoBol Drashek EG Frank von DRashek, Rochester Technology Group. Code AKie, wasn't going to be telling much of anything to Anyone. Lips sealed, ears removed, Eyes, Polarized from Light of public Need. Suffragettes....
So public slowly crawled forward & enough of myself Remains to say, told You.(did have to learn to write again) HTML preceded cobol, as its studio display language. by 1960 languages where sprouting like Victory Gardens.
So Where Thomas Moores Law Now? bTW Moores Are Mafia Resturantuer Family From Edina, Village founded by City of Edinburgh, Scotland in 1845. MOORES Manufacturer Soda Pop. NOT Electrical Engineering, so much.Yet, In homi Way, Listen To ALL Those Blabblers.
Well Size of Trannie started coming into reality of usefulness this decade, so trannie count to mimick needed hardware is in order & todays' Moores law is Trannie Count. Something going UP Faster than Volcano in heat. NEED IS GREAT.
Next, When Trannie count Is High, Estimated to be 8 billion+, by Ultee', IS MATRIX & SSD, Back to 1960s', With Finalized Hardware.... Moore hardware IS Needed To Further Develope Moore Matrixes. Hardware needs work with old hardware & adopt new standards as well. so its ever increaseing need. With SSD theres NO reason 16 bit games cann't be revived & played. Only in HardWired CPU of Today is bit width & string lenght Important due to HardWire. In SSd it can be DYNAMIC.
So whats cpu thats cable box, game machine, game card, multi O?S beast to do, except develope More Matrixs to run.
Matrix & memory design itself become Moores Law, increaseing till, well, No Moore Useful events can be Softwared into its OPEN Architecture. Start Your Own Telco, Matrix. Start Your Own Radio/Tv station. Start Your Own CAD/CAM. List Just Staggers Mind. yet SSD isn't limited to small size chip, its software that makes it HUM. Software Doubles in Size, Matrix takes more space to run, Need Bigger Matrix. 64Gb Matrix. USB Thumb Matrix. Plug IT In 'Ur Ear? Well, Maybe NOT. Medical Matrixes,O.K.. DMV Matrix. Even Computer 'Zine Matrix. Remember OLD Days When Computing was Done on Computers?Hahaha. Well You Get Picture. Arrays as Small as Terrabyte block? Wow, My, Gosh....What will think of Next.
Anything that is predicted or perceived to be able to continue "forever" will always come to and end before the hopeful optimists (or vested interest) predict. History has proven that countless times.
The electronics revolution has been one of the greatest technological achievements of mankind, and it will clearly continue for some time. Whether cost factors or technology barriers will ultimately end it is currently not known, but it is very safe to predict its end, the only difficult question is when.
Memristor proliferation may actually accelerate integrated device denisty past Moore's Law for a few years.
Moore's law may have to be changed a little bit, from "transistors" to "switching devices" or something.
Point is we're only getting started
The only reason why the manufacturing process has followed Moore's law is because people TRY to make it follow Moore's law.
If Moore had fudged the math and come up with a 1.5 increase every 2 years, or a 2-fold increase every four years, then that would be what the industry follows.
Another self-fufilling prophecy: I predict you will suspect your wife (or girlfriend or that woman you stare at but never have the nerve to talk to) is a cheeting whore.
This is all wrong!
When gears of the primitive mechanical computer couldn't be shrank anymore, we invented vacuum tubes. And when the vacuum tubes did hit the shrinking wall as well, we invented the transistors and then integrated circuits. There are plenty of emerging alternative to the IC transistors. Apparently, the next big thing is 3D self organized molecular circuit. How about quantum computing? We are just shifting from one paradigm to another. There is no wall. There never was and probably never will.
Take a look at this excellent presentation about this subject:
http://www.ted.com/talks/ray_kurzweil_on_how_technology_will_transform_us.html
Ramon
Intel's own scientists have also predicted the end of Moore's law:
http://news.zdnet.com/2100-9584_22-133066.html
One should note that all features do not have to scale down similarly with process technology, so the will no doubt be a couple of process nodes after 16 nm.
http://en.wikipedia.org/wiki/16_nm
All the technologies you speak of involve replacing something bigger with something smaller. The wall is the size of the atom, you are not avoiding that by switching to molecular circuits or similar...