Photolithography is limited by the wavelength of light. Â Below 160 nanometers, UV is not light, but ionizing radiation. Â So at some point, have to switch from photolithography to contact lithography, such as imprint lithography, or direct contact printing. Â But did they stop shrinking stuff before we reached that limit? That they are lying about it suggests moral decay, rather than exhaustion of the low hanging fruit.
Current photolithography is stuck using UV at 193 nanometers. Â Could they have gone lower?
Nitrogen and fused quartz, which they are already using, is good down to 160 nanometers. Â They are using water for their high refractive index fluid, and water becomes opaque below 193. Â Perfluorocarbons, however are good all the way down to 160 nanometers. Â So they did not push current technology to its final limit. Â It was not exhaustion of the low hanging fruit that caused the current alarmingly indefinite pause in Moore’s law.
My prediction is that if humans resume technological advance, will resume in China, and will resume using contact lithography or direct contact printing, as there is little point in recreating from scratch a technology that has reached its ultimate limit. Â And if humans don’t resume technological and scientific advance, it will be a long, slow, painful and messy descent into a dark age, until harsh conditions cause natural selection to resume.
People who prattle about exusting the “low hanging fruit” always fall into deny the decline camp. During the scientific revolution the west very consistently went off in a different direction and found new areas to exploit our tech whenever we hit a wall in one area. Hell we didn’t discover how the Roman made their concrete until a few years ago and we built the entire modern world with the vastly inferior Portland concrete by continually finding ways around it’s limitations.
http://newscenter.lbl.gov/news-releases/2013/06/04/roman-concrete/
Goes even farther back than the Romans. Here’s a quote form a page on MgO cement.
“…China is using MgO cement almost exclusively now. They are restoring ancient temples that were made with MgO formulations 2500 years ago. Some of the ancient columns he has worked on have a smooth finish, almost indistinguishable from marble. While investigating the ancient technique, they chipped some away and found the wood columns inside were perfectly preserved…”
http://www.instructables.com/community/Ever-heard-of-Ceramicrete-MgO-cement-Magnesium-c/
Even more…this chemist figured out a lot of the pyramid blocks (upper levels, corridors and outside finish) were actually a form of concrete called geopolymers. Perfectly explains the tight fit between blocks. Egyptologists say he’s wrong but I’ve read all his books and his case is so tight he must be correct.
Books
http://www.geopolymer.org/learning
Website-plenty of papers if your interested.
http://www.geopolymer.org/
Geopolymers also explains the stone structures in South America with irregular close fitting stones.
Will the descent really be that slow? It has just started, but is already obvious.
Our technology isn’t declining, just advancing more slowly than it used to, perhaps asymptotically. So why must our society decline?
Oh right, because we divert an exponentially increasing flow of wealth from productive investment to needy voters, whether or not our technology can support this. Hugo Chavez pillaged the oil industry to pay for welfare programs, so that his country now has neither.
My prediction is that technology will continue to advance, albeit slowly, in undemocratic city-states like Singapore, where people who refuse to contribute will be kicked out. The Republic of Venice rose from the ashes of the Roman Empire and survived 1100 years.
Our morals were always and will be always ready to decline nearly overnight.
We just saw it happen. It’s called 1965 on. All that is necessary is for a lack of will among elite men. Possibly connected to allowing academic deferments from conscription during World War II. It doesn’t actually mean your fighting is done by fools, or we’d lose wars on the battlefield. It does however mean your thinking is done by cowards and shirkers, and the longer the Enlightenment went on the more thinking was rated over courage and work.
Our problems are in our elites, and only with their destruction can we begin to build again.
China won’t save humanity, it’s not in them.
And if you’re not Chinese, they won’t consider saving you.
Intel released tri-gate based Ivy Bridge chips in 2012; after announcing the technology in 2002. Global Foundries and TSMC claim to be developing similar technologies for real soon now; Intel has always used their process advantage to keep AMD out of the high end.
Transistor counts are still going up http://en.wikipedia.org/wiki/Transistor_count
Intel and Global Foundries are based in the US; though Global Foundries, once AMD’s maunfacturing business, was sold to Abu Dhabi. TSMC stands for Taiwan Semiconductor Manufacturing.
That graph ends in 2008. The last reduction lithographic scale, possibly the last ever reduction in lithographic scale, occurred in 2007.
However, transistor counts are still going up – but chip area is going up similarly. A more interesting metric is the cost per transistor per hertz, which is no longer going down.
Since transistors and their connects are not getting smaller, increased transistor counts are accommodated by bigger chips, and by vertical integration on chip, which does not reduce price, decrease power consumption, or increase speed.
It sounds like we’ve hit the natural limits of silicon (or close to it). The real test will be if we found a way to side step it or if we just make bigger chips with more and more marketing.
The natural limit of our current photolithography technology s 160 nanometer light. We are not there yet.
We are a long way from the natural limits of silicon transistors. doped gates are approaching their natural limit, but gate size is far from being what occupies most of the space on chip, and we have in any case largely given up on doped gates for undoped gates, in part because doped gates are approaching their natural limit.
And, when one approaches the very small, lots of non transistor technologies for switching become available.
Taiwan-based TSMC and Global Foundries, owned by Abu Dhabi, most fabs in Singapore, compete with Intel, which uses its near-monopoly on high-end processors to buy more advanced fabs, thereby shutting TSMC and GloFo out.
At least, that’s the story people tell themselves. Is it wrong? Why does Haswell perform better than Sandy Bridge or Nehalem? Does Intel no longer have an advantage in manufacturing; and their chips perform better because they have better design than AMD now?
Most of our phones use chips built in Taiwan and designed in Korea, so looks to me that intel and TSMC are running neck and neck.
that’s because phones use ARM chips, which Intel hasn’t produced in years. Intel’s strategy for the mobile market has been to make efficient x86, which they claim to finally be achieving with Bay Trail. Well, there are certainly lots of tablets and netbooks on the market now with Bay Trail parts.
TSMC, otoh, makes AMD’s current Kabini and Temash chips. I’m using a netbook with a Brazon chip, the predecessor to Temash, manufactured by Global Foundries.
AMD claims that their current Temash chips beat Bay Trail. I’m waiting until Mullins to upgrade, though, which rumor has it will be based on TSMC’s “20nm process”.
Playstation 4 and XBox One are based on TSMC “28nm” chips. Meanwhile, AMD’s high-end Kaveri APUs use Global Foundries “28nm” process. TSMC and Global Foundries have had similar capabilities for some time; Intel claims to be moving to “14nm” in the fourth quarter of this year (they were claiming that Broadwell would be released by now, but I guess they missed their target).
Since Intel only fabs Intel chips, and Global Foundries was until recently AMD’s fabs, TSMC currently makes most of the smartphone chips. Rumor has it, Apple is looking to contract some smartphone and tablet chips from Global Foundries.
Phones, like computers running Linux and unlike computers running windows, are not bound to any particular CPU architecture. Thus phones are a better indicator of whose technology leads than desktop computers running windows.
You can compare an Exynos-based Chromebook with the Bay Trail netbook right next to it next time you visit a computer store.
http://anandtech.com/show/7428/asus-transformer-book-t100-review/4 compares a Bay Trail tablet to a bunch of ARM-based devices.
Intel hasn’t been able to scale x86 down to fit in a phone yet. Some say it’s never going to happen; I don’t know enough to have an opinion on the matter.
Anyway, Samsung makes their own Exynos chips and have been making chips for Apple too. Some of them may even be fabbed in Texas. My friend I was just hanging out with has a Nexus 7, with a Qualcomm-designed chip that might have been manufactured by either Samsung or UMC (another Taiwanese company).
It is most unfortunate that American conservatives are so strongly anti-science. The party of Jesus’s resurrection on the third day won’t be the party of digital mind reconstruction research.
meanwhile, just last night, Barack Obama introduced Dr. Neil deGrasse Tyson’s new show, a reboot of Cosmos. It was about how the evil Catholic Church persecuted scientists, by focusing on a totally irrelevant guy that Cosmos sheepishly admitted was not actually a scientist, and quietly admitted was actually a heretic (a Universalist).
Any decent astronomer would know that this is misdirection, and refuse to narrate the program. Dr. Neil deGrasse Tyson is a professonal Black man.
Who thinks that CO? emissions are the worst thing ever and radiation is too scary to contemplate? Who thinks that creationism is ridiculous and humanity’s story begins with Monkey Buddha being enlightened by reason, choosing xir race, sex, national origin, sexual orientation, gender identity, and becoming the first human?
But yeah, it’s the conservatives who are so strongly anti-science. We know this because they refuse to accept the science of gay marriage and transgenderism.
Anti official science. These days anyone who attempts to do replication is deemed anti science.
Peer Review is consensus, and consensus is the opposite of science.
If you pay attention to the views of a bunch of people who are paying attention to external reality, to the empirical data, and take the median or mode of their views, you will probably get pretty close to the truth, but if those people are not paying attention to external reality, but to each other’s views, which is to say, the scientific consensus, you are hosed.
I am a university student and in one of the early courses I had to take, the students were divided in groups. We had to do a work about a subject using scientific articles given by the teacher. One of the groups found by themselves that two articles (I believe by the same person) that they were using, had wrong data.
I doubt that the exhaustion of the low hanging fruit theory explain why I don’t have a flying car in 2014 and the US don’t have any manned space craft
Flying cars don’t exist. Helicopters, airplanes, and blimps do; helicopters and airplanes are hard to fly and blimps are slow. Perhaps the future can involve robotically controlled personal flying machines; but robot cars will work a lot better, and happen sooner. Just because something would be cool doesn’t mean it’s a good idea.
However, with respect to manned spacercraft, this really is a political issue. NASA doesn’t have the budget to hire a quota of Black engineers to send a lesbian Hispanic Muslim to Mars. They try to use contractors and subcontractors to weasel their way out of affirmative action requirements, but every time a middleman is used the quality goes down and the cost goes up.
Flying cars are trivial given cheap energy and high-energy density storage. Energy use per capita and energy technology stopped growing in the 70s, when everything else stopped improving as well. We could have had flying cars by now had the trends continued.
Still, flying cars probably suck; most trips are fixed route, and the biggest problem is traffic and speed. The more concerning thing is that we can’t even build high speed rail transport, which neatly solves all of that.
yes, yes, and day trips to the moon are also trivial given cheap energy and high-density energy storage.
the biggest problem with flying cars is that people will not maintain them, and they will crash, and they will die, and they will kill others.
Trips to the moon are not physically plausible without nuclear rockets. You could run a flying car on kerosene, though.
Jim, this is really funny. I just spent an enjoyable half hour not working. thanks.
http://esr.ibiblio.org/?p=5238&cpage=1#comment-428965
That sounds like I’m belittling your efforts. It’s also very good. you provide some really interesting specific rebuttals I haven’t seen elsewhere. If others haven’t perused this thread, you should.
Perhaps foundries have not shifted to 160nm light because all they hopes where pinned on 10nm light, but they have not manged to get light sources bright enough to work after the mirrors have absorbed most of the light. So the stopgap has been multi patterning, not really sure why they have not shifted to 160 nm light as multi patterning is expensive. Superlenses may become used as than can also push resolution down without changing wavelength. My bet is the next method will be physical contact stamping, or perhaps large arrays of compact electron beams.
Here is an interesting aproach:
http://www.smartplanet.com/blog/science-scope/the-next-generation-computer-chips-might-be-made-of-dna/1760
If they can make a unit of a FPGA the you can simply crystallize out of solution as big a computer as you want.
Jim don’t end up a fool like every Malthusian before you. Meanwhile we have a serious threat of an imminent Dark Age. Let’s understand the distinction.
Whoops.
[…] Of course all this crap comes from not understanding what actually drives Moore’s Law. Transistors don’t make themselves after all. They are developed through a quite complex process, which I don’t presume to understand. Jim appears to understand it though, and he says Moore’s Law is dying. […]