What the History of Computing tells us about their Future MMX 2010 MXXVI 1026 Better ideas "There is no reason anyone would want a computer in their home." -- Ken Olson, president, chairman and founder of Digital Equipment Corp., 1977 to the World Future Society Mainframes Big Slow and Dumb More Powerful Bit width plus 4 8 16 32 64 Transistor count Co-processors math graphics heart problems Faster Clock Speeds http://en.wikipedia.org/wiki/List_of_device_bandwidths Hold more Chip Feature size Smaller Glossary (do one as a handout) Government and corporation, department and workgroup, family to personal Mainframe mini video resolution pixel size Physical limits http://freakonomics.blogs.nytimes.com/2008/04/10/our-daily-bleg-does-640k-really-belong-to-bill-gates/ Survey of news reading habits http://people-press.org/report/652/ Tracking of IP numbers (also MAC numbers - like barcode, for computer) In 1996, the average speed of an internet connection was 50Kbps. In 2010 the average speed of a broadband connection is 3.9Mbps - an increase of over 74 times. In theory, this should mean that the internet is 74 times faster than it was in 1996. The current Whitehouse.gov webpage has grown from 22 KB in 1996 to 1.2 MB in 2010 - an increase of 54 times. With the effects of latency and distance, the load times of the page are about the same in 1996 as they are in 2010. http://www.readwriteweb.com/archives/a_faster_web_-_its_not_about_the_network_anymore.php QR readers do example? GPS tracking of walk feed into medical reports Phone clock and timer microphone camera compass accelerometer GPS Fight between phone carriers and internet Labels on file folders? All but 6% used Internet IP numbers IPv4 vs IPv6 32 bit vs 128 bit 340,282,366,920,938,000,000,000,000,000,000,000,000 unique IP addresses 340,282,366,920,938,463,463,374,607,431,768,211,456 Three hundred and forty undecillion, two hundred and eighty-two decillion, three hundred and sixty-six nonillion, nine hundred and twenty octillion, nine hundred and thirty-eight septillion, four hundred and sixty-three sextillion, four hundred and sixty-three quintillion, three hundred and seventy-four quadrillion, six hundred and seven trillion, four hundred and thirty-one billion, seven hundred and sixty-eight million, two hundred and eleven thousand, four hundred and fifty-six. That’s a big number. http://itknowledgeexchange.techtarget.com/whatis/ipv6-addresses-how-many-is-that-in-numbers/ Garbage weighed, charged by weight. Five billion mobile phone subscriptions http://www.readwriteweb.com/archives/global_wireless_subscriptions_to_reach_5_billion_in_september.php Most people could connect to the internet, but they can not afford to do so. 35 billion devices connected to internet already Things that know where they are - aircraft, ships, delivery vans, taxis Use cell phone camera to measure smog via crowd sourcing. Augmented reality Learning - Blackboard Worldboard point at any object, get told what it is. Example of music via that iPhone app? Name of app? http://www.worldboard.org/pub/spohrer/wbconcept/default.html If objects could talk http://blog.michaelmassie.com/post/1093498539 In two years, 90% of all data will be video http://www.readwriteweb.com/archives/cisco_futurist_predicts_internet_of_things_1000_co.php A central nervous system CeNSE for the earth, with a trillion sensors Thumbtack sized sensors on bridges http://www.readwriteweb.com/archives/cense_hp_labs.php http://www.hpl.hp.com/research/quantum_systems/ http://www.hpl.hp.com/news/2009/oct-dec/cense.html The panopticon is coming imagine a future where all objects are "social" data-collectors who can report their use, their history, their location, etc. Now imagine the government or corporations accessing that data and taking action based on what the objects' data tells them. Goodle Earth Google Maps Street View stops at our entrance 2003 - As suppliers continue to ramp up production numbers the price of blank 12-inch silicon wafers continues to fall. Early in the year 12-inch blank wafers were about US$500, but have since fallen to $200, reflecting the nearly three-fold production increases by some suppliers. The up-turn in the I.T. industry as a whole, and specifically the recovery the semiconductor industry has seen, is a large contributing factor to the increased demand and production numbers. Though the industry is shifting to 12-inch wafers, it appears 8-inch wafers are still seeing increasing sales numbers, with continued high sales being projected through the first part of 2004. With 8-inch wafers currently selling for US$30-40, suppliers are considering raising prices, which will further reduce the gap between 8- and 12-inch wafers, which would in turn help speed the industry shift to 12-inch wafers. Wafer manufacturers don't want to spend the money to maintain production of multiple size wafers if it isn't necessary; this costs money that takes away from their bottom line. pie * (diameter/2)^2 = area (2:21pm EST Fri Dec 05 2003) 8 and 12 inches refers to the wafer diameter… silicon area cost for 8 inch wafer: $40 / (pie * (8/2)^2) = 2.5/pie silicon area cost for 12 inch wafer: $200 / (pie * (12/2)^2) = 5.5/pie Here is some real math: A 8 In. wafer has 50 Sq. In. to be used for chips at $40 that equals $0.80 per Sq. In. A 12 In. wafer has 113 Sq. In. to be used for chips at $200 thats only $1.77 per Sq In. only 17 cents more than 2 times the amount. also you have yields Its cheaper to handle and work with one wafer than two when your design is make and exactly copy but when you dont exactly copy your yeilds can suffer if just a few bad wafers occur - by Nataku Frankly, it's a shame that you didn't get good answers. The prices shown above are for RAW Wafers (PURE SINGLE-CRYSTAL SILICON WAFERS) about 1/32nd of an inch thick [800um]. Intel & AMD don't make wafers – they use wafers made by companies like S.E.H (Shin-Etsu Handotai), Siltronic (formerly Wacker, GY), SUMCO (Sumitomo/Mitsubishi – Joint Venture), and MEMC (formerly Montesano Electronic Materials Co. – current stock symbol WFR) — by the way, if you had purchased WFR about a month ago, you would have already made about 25% increase……….. SUMCO issued in November the largest IPO this year on Japan's stock market – USD $1.3 Billion Dollars (it saw an 18% jump on first trading day in Japan). As for 24″ wafers — perhaps someday, but they are not in the current forecast. We made a product for a Japanese research lab about 5 years ago which would make the raw wafers at 400mm (15.75″) — but current thinking is that the next major change could result in 450mm silicon wafers (~18″ diameter) — but like the 12″ which have taken about 8-10 years to get into the mainstream (12″ still are behind 8″ in total square inches) it will take another 8-10 years (and BILLIONS of DOLLARS) to establish a new standard. Calculations shown by the others aren't exactly correct since actual wafer diameters are in millimeters (200.0 & 300.0 mm), so actual area calculated in Square Inches is 15.500 (200mm-8″) & 34.875 (300mm-12″) — so the apparent next generation size will be 450mm with an area in Square Inches of 78.469. As you can see in this progression, each new wafer size recently adopted has more than doubled the useable area for chip manufacturers. For example, a 6-inch (150mm) wafer has a surface area of approximately 27.4 square inches, whereas an 8-inch (200mm) wafer has a surface area of approximately 48.7 square inches. Thus, the 8-inch wafer has approximately 78% more surface area than the 6-inch wafer. A 12-inch (300mm) wafer has a surface area of approximately 109.6 square inches or approximately 125% more surface area than an 8-inch wafer. Despite the industry's focus on larger diameter wafers, we continue to manufacture and sell a significant amount of 4-inch (100mm), 5-inch (125mm), and 6-inch (150mm) wafers. ODDLY enough, in addition to tight supplies of 12″ wafers, certain 4″, 5″, and 6″ wafers are also in short supply as some wafer manufacturers have moved all production to 8″ or 12″ wafers. Another factor affecting cost of Silicon Wafers is the short supply of the raw material – polysilicon (pure silicon, but not single-crystal) — this comes due to a huge increase in demand for solar-powered devices last year. Did you buy any solar lights to put next to your sidewalk? Many people did. New polysilicon capacity won't be online for a couple of years… 2010 - Joo-Tai Moon, senior vice president of Samsung Electronics, said that the industry will move to 18-inch wafer technology by 2015. The 12-inch, 300mm wafers used today can yield 2.25 times more chips per wafer than the older 8-inch, 200mm wafers, yet they take just about the same amount of time to pass through a factory, reducing the cost per chip and significantly boosting total monthly output. An 18-inch wafer plant would show a similar reduction in per-chip cost and increase in output. It has been estimated that a factory designed to make chips on 18-inch wafers could cost between $12 billion and $15 billion to build, nearly triple the price of an equivalent 12-inch wafer factory. According to Digitimes, Moon believes phase-change RAM (PRAM), oxide-based memory and spin-torque-transfer magnetic-random-access-memory (STT-MRAM) will be the stars of tomorrow. These next-generation technologies could all be seen as high-efficiency and low-cost alternatives to today's memory standard. Samsung is spending $7 billion upgrading its plants with a large chunk going to upgrading to 30nm process technology. µ 2008 - Polysilicon is the black material that is most visible when one looks at a microchip. A chip's circuitry, however, is made with monocrystalline silicon, which is polysilicon that has been crushed, melted, and grown into ingots with a singular crystal orientation. Polysilicon is also used—in monocrystalline form by some manufacturers—to capture the sun's energy on most of the photovoltaic solar cells that are sold today. Until a few years ago, demand for polysilicon from the solar industry was so low that it could be supplied from piles of scrap material deemed too impure to be sold to semiconductor manufacturers. Not anymore. The solar industry uses polysilicon of a purity level of 99.9999999% (seven 9s), whereas semiconductors require material that is at least nine-9s pure, he adds. It's difficult to get an exact idea of how the supply shortfall has affected the price of polysilicon. Winegarner says the spot price of polysilicon can reach as high as $500 per kg, which is more than 10 times what the material costs to make. But prices on the spot market are misleading, he says, because almost all polysilicon is sold through long-term contracts. Hemlock's Doornbos says he can't discuss pricing. It's clear, however, that polysilicon is an expensive material and cell manufacturers are keen to find ways to consume less of it. The main tactic so far is to use thinner polysilicon wafers. Mostly through process improvements in polysilicon slicing, solar wafers are now 180 µm thick, down from 280 µm a few years ago, Winegarner says. Approximately 10 gms of silicon is needed to build modules capable of generating a watt of electricity under the strongest sun. Polysilicon is expensive and the modules thus built cost up to $ 4.50 per watt of power generated. At present, there is a polysilicon shortage in the world which has raised their price from $ 40 per kg to $ 100 per kg even for long term contracts. A new polysilicon plant of annual capacity of 2,000 T py (good for modules that will generate 200 MW of power) would cost $200-300 million and take two to three years to build and stabilize production. 2002 - move to 8 inch But the 50% increase in diameter yields 2.5 times more surface area for etching chips, yet costs only about 20% more to process. While some existing 200-mm fabs will continue in production for a decade, they'll have a virtual cap on revenues. If any market served by a 200-mm producer gets large enough to attract a 300-mm rival, "bang--the game is over," says Trevor Yancey, vice-president for technology at market researcher IC Insights Inc. "A 200-mm producer simply can't compete on price." Since building a new 300-mm fab won't be feasible for midsize producers, Yancey adds, "the only real choice may be to go fabless"--farming out production to a foundry. Area is pi r squared 100 mm 7 854 150 mm 17 671 200 mm 31 416 300 mm 70 686 450 mm 159 043 2002 - The price of a chip factory, or wafer fab, that can handle 300-mm platters can easily top $2.5 billion--and some plants are pushing past $3.5 billion. Only a handful of traditional chipmakers can justify such an investment because keeping one of these "megafabs" humming requires annual sales of at least $6 billion. By 2007, as many as 40 megafabs may be up and running, with a combined capacity of 575,000 wafers per month. The 12-inch (300mm) wafers used today, can yield 2.25 times more chips per wafer than older 8-inch (200mm) wafers, yet take just about the same amount of time to pass through a factory, significantly boosting total monthly output. silicon consumption increases with a long-range compound annual growth rate (CAGR) of approximately 10%, which equates to a doubling in volume about every 7.4 years. When the wafer area increases by >2 times, but the cost of the new tool set for the same number of wafer starts increases by only 30-40% (which is typical), the cost per area decreases by 30-50% -- an annualized improvement of ~4% when wafer size changes occur about every 10 years When cumulative margin in fig 15 exceeds cumulative spending the investment starts to have a positive payback. Using the assumptions of fig 14 the payback time could range from 3-8 years with a more probable range of 4-6 years. This is assuming that the required market volume can be met with a total investment of $1.3B and $2.6B. While this situation is obviously oversimplified one can start to see the difficulty for wafer manufacturers. Many factors can erode margins in any business. In the case of 450mm wafer manufacturing we are pushing into the realm of higher yield losses relative to smaller diameters and heavy investments in order to meet the technical challenges. The risk of such a venture is considerable. http://www.itrs.net/Links/2009ITRS/2009Chapters_2009Tables/2009_ExecSum.pdf Well, I would have thought it would have been bigger, but Uptrends.com, a global webs site monitoring company, says the 1024×768 screen resolutions size is used more often than any other size. The 1024×768 screen resolution size has the largest share of global usage, with global usage of 25.17%. But less and less people are using this resolution. In April 2007, the share of global usage was 55.34%. Internet users worldwide have been increasingly choosing larger screens with higher screen resolution. Apple might make a note of this, as the company seems to think that the world wants increasingly smaller displays (witness the latest iPad nano). In a study conducted by Uptrends, the usage of smaller screen resolutions such as 800×600 is also on the decline. For example in April 2007, the global usage of the 800×600 screen resolution size was 8.18 percent. In this year’s study, global usage for this resolution is down to 1.70 percent. In August, this is how screen resolutions stacked up: 1024×768, 2.517%; 1280×800, 17.93%; 1440×900, 9.81%; 1680×1050, 6.71%; 1366×768, 5.20%; and 1920×1200, 4.17%. I predict the 1920×1200 screen resolution will grow significantly in popularity, and that we’ll see more and more 2560×1440 (a la the 27-inch iMac) as folks increasingly use their computers to double as entertainment devices such as TVs and movie viewers.