You can't patent the effect, but a USPTO search gives 77 patents on an assortment of devices (including one suggesting that you should sputter radon atoms into the disc surface - holy radioactive storage, Batman!) exploiting it, and US patent 6441661 (assigned to Fujitsu) looks as if it's on GMR magnetic sensors in general.
Does anyone have tools for traversing the graph of patents under reference in both directions? Key patents would tend to show up at the top of lists sorted by number of citations, but I don't think that's something I can sort on in the USPTO database.
That the words 'texture', 'instruction pointer' or 'blitter' appear nowhere in either PDF file is a bit of a giveaway.
Whilst the registers are essential for getting any kind of driver to work, the documents don't describe the exciting features of the graphics processor. They give you enough control over the memory-controller timings to convert any Radeon card into a smoking brick with a small kernel-mode driver, but they don't give instructions which actually make the graphics silicon do things. There's no indication of what the machine-code for the vector processors looks like.
If you compare this to the documentation that Intel has for its (obsolete) 845 graphics controller, you notice that the whole block of registers for controlling even something as basic as the blitter, let alone the 'set instruction pointer for processing unit N' registers which actually let you set the high-performance processing units in the card to work, are missing.
These documents let you use an R500 or R600 card as a frame buffer. Not worth making a song and dance about that one.
Myself, I'd be fascinated to see documentation for the Intel G965 like the documentation for the G845; it clearly exists, there's a paper in the most recent Intel Technical Journal about low-level programming on the 965, it's just not available to mortals unless by attempting to reverse-engineer the x.org 965 driver.
This was just a scheduling problem; to figure out the orbit of Dysnomia, you need to take several exposures at different times using a telescope facility capable of resolving Eris and Dysnomia. Which means either the Keck laser guide-star system, or Hubble. Each of which has a queue of astronomers reaching three times round the block who want to use them; and the queueing systems are designed for 'I want to spend a night taking spectra of this radio-galaxy' rather than 'I need pictures which take a five minute exposure, but I'd like one every 30 hours for a month'.
Also, it takes ages to get papers written and published; the Hubble pictures of Dysnomia were from August 2006 and the paper (in Science, a journal which is really quite quick to publish) came out yesterday.
The 'special crystal which took three years to grow in Russia' that the article writes about is made of pure silicon-28 precisely to get around this objection.
I think the concern is that samples of silicon from different sources (consider, for example, 'depleted silicon' from the scrapyard of the Russian isotope-enrichment facility) might have different isotope distributions at the 10^-7 level, whilst good laser enrichment can ensure a really very constant isotope distribution.
For larger-scale effects, the Miravi site has some very nice data, though I'll admit it's not as well-presented as Google Earth. It's from the MERIS instrument on the European Space Agency's Envisat, which takes a constant thousand-kilometre-wide swathe at 250 metres per pixel as it heads round the Earth in a 101-minute orbit, retracing its steps exactly every 35 days.
You can watch reservoirs filling, watch rainforest clearance in Brazil (though not so well in Borneo, since the country is almost perpetually clouded over), look at algal blooms in the North Atlantic, and see the smoke from volcanic eruptions and big fires. Over a few years you might be able to watch cities grow in China, though it's surprising how well cities blend into the landscape at 250 metres per pixel.
The data's available as JPEG files, and the only slight problem is that Firefox regards JPEG files of more than 2^15 rows as corrupt.
Romania is probably the hardest country of Eastern Europe to intimidate by means of gas supplies; it has quite substantial local production of oil (Ploesti used to be the oil capital of Europe) and of natural gas, a couple of modern nuclear reactors at Cernavoda on the Black Sea coast, and exports electricity.
Central Romania feels very energy-poor, but that's an infrastructure rather than an availability issue; it's a big place, and not a wealthy one, and they haven't yet got round to putting in the wires and the pipes universally.
Large memory modules mean otherwise-unavailable large memory capacity in a single system, since systems have only a fixed number of memory slots; top-end CPUs give you single-threaded performance which is otherwise unavailable.
But very few large-data-storage applications are heavily constrained by the number of drives in the system, so almost always you can use two small drives instead of one drive of twice the size; yes, more drives take up more space, but very rarely is the space taken up by one drive going to cost more than twice the price of the disc drive.
I could see '2.5" drive of unprecedented capacity' as a useful provision of an otherwise-unavailable facility, since laptops tend to have exactly one drive. But there are remarkably few systems that are constrained to have only one 3.5" drive.
With the ready availability of RAID cards and drivers, I wonder what the point of a terabyte drive at well above the current best cents-per-gigabyte pricing is.
This costs three times as much as a 500G drive for twice the capacity; all the reviews are pointing out that you'd want to run two drives in RAID1 because 900G of data loss from a single mechanical failure is unlikely to be acceptable; six 500G drives in RAID6 get you twice the capacity, more reliably, for the same price, three 500G drives in RAID5 get you the same capacity, similar reliability, half the price.
Yes, six drives are a little louder that two drives, but drives have been quieter than CPU/PSU fans for a long time now. I suppose some cases don't fit six drives very conveniently, but you can get quite a fancy case for $400.
I suppose there might be hosting situations in which the volume of two hard drives is worth $400.
Yes, you can make various hot-ices on Earth using a diamond-anvil cell - the pressures required tend to be in the gigapascals, so tens of thousands of atmospheres. At fairly high pressure you get ice-7, where the oxygen atoms in the water form two interpenetrating cubic lattices; at higher pressure you get ice-10 where the oxygen atoms are all lined up in a single cubic lattice.
The whole point of this patent is that you can't take the drive out and copy it, because you can't send ATA commands to the drive without authenticating yourself, and you can't authenticate yourself unless you demonstrate your ability to compute (reading between the lines of the patent) SHA1(challenge ^ magic_number_stored_in_Tivo_firmware).
Like all of this kind of authentication processes, I expect this one to last until someone gets fed up, reads off the drive firmware from the ROM chip on the board, and patches it not to do the checks. Though if Tivo have actually convinced WD to take the enormous debuggability hit from embedding the ROM in the drive-controller ASIC rather than having it as an externally-accessable device on the board, this is not a bad security approach.
That is a dreadful patent, and it would be ridiculous to see it issued; hardware challenge-response dates back to at least the first IFF machines in the second world war, they're not even mentioning having a deliberately slow password-hashing algorithm, which is itself at least as old as UNIX, and the technique is vulnerable to bump-in-the-ATA-cable extraction of the data from the disc in the first place, and probably also to an attack where you swap the drive controller board for one from a drive of similar model without Special Tivo Sauce.
I can't be the only person whose reaction to the article is to google 'David Mullins', and discover that it's a reasonably common name, shared by the professor of housing policy at Birmingham University, the director of academic administration at Warwick, a 1991 Stanford math grad student, a London-based artist, and the ex-vice-chairman of the Federal Reserve Bank of Minneapolis.
I think it's distinctly unprofessional for an organisation to record the fact that it fired someone in a document on the Google-accessible web, since having the fact available without the explanatory information that you'd get from asking the organisation for a reference might well be prejudicial to future employment for the person concerned.
NGIA is the map-making part of what used to be the National Reconnaissance Organisation, the consumers of vast amounts of black-budget money.
There are all sorts of censorships in Google Earth, from the glaring (there are no roads or cities in Israel!) to the moderately glaring (you can count planes at Beirut airport, not so at Ben Gurion) to the subtle; I had a friend out at the RAF base in Basra a few months ago, and was a little alarmed that there was one-metre georeferenced imagery of the camp available - though since the camp's still there and the impacts on it seemed fairly random, it would appear that the local insurgents didn't have GPS-guided mortar rounds. Or possibly GPS was jammed over the camp -- after all, the airmen in the camp know where it is already -- though maintaining the navigation hardware in planes when you can't test it in-base would not be fun.
That made the British news, and maps.google.co.uk now has no GIS information for Iraq at all (in fact, no GIS from the Egyptian border to the Indian border), though it still has sub-metre imagery of central Basra
Noordwijk is the home to ESA Mission Control, but that shows up on the map without trouble. I'm by a long way not interested enough to spend a hundred Euros getting myself to Holland, walking down Albert Verweystraat taking photos of the buildings, and seeing which one belies the fact that it's labelled National Marine Conservancy by the platoon of marines outside.
Don't buy 60-watt-equivalent bulbs, then. If you're concerned about brightness, buy the brightest bulbs available, they'll still use way less than 60 watts of power, and if the peak light output is 150-watt-equivalent then the brightness during the warm-up period is still usually more than the 60-watt bulb you started with.
OK, I suffer somewhat from SAD in the winter, so opt to have as much light around as possible.
Certainly there are urban legends around which are propagated by conservative propaganda sites.
But that there are five milligrams of Hg in a compact-fluorescent lightbulb is not one of them; in particular, the link that you provided admits that.
I too have a house full of CFLs - people complaining that 60-watt-equivalent CFLs are too dim are taking slightly the wrong approach, CFLs are so much more efficient than incandescent lights that you can put, into a fitting that can only handle 60 watts of heat, a 23-watt CFL which is equivalent to a 150-watt incandescent. My study is lit with three 23-watt CFLs, which provides a really excellent reading light... with the low power consumption, you can use cheapest-available desk lamps to put the bulbs in, and place them wherever's convenient.
The permitted mercury content of tuna in Canada is 0.5 part per million, so a 170-gram tin contains at most 85 micrograms of the stuff, about a factor sixty less than the lightbulb.
I think this is more a story about how good we are at detecting minuscule quantities of material, and how political requirements tend to be of the form that the allowable amount of a dangerous material should be a small multiple of the detection limit; I would wager that the health damage caused by the stress of being told to find two thousand dollars to decontaminate your living room is significantly greater than any that could possibly be caused by five milligrams of mercury vapour.
This is because part of the point of the article is that this new equipment reduces enormously the corona-discharge losses which are producing the amazing sparks... marvelling at the sparks in that picture is similar to marvelling at how loud your car engine is; every joule spent in sparking is one that's not being put into creating the fusion environment.
This isn't where I would have expected IBM to put Cells; from the first announcements several years ago of Cells and of the Blue Gene architecture, everyone's asked 'when do we get a Blue Gene made of Cells?'
That may be something that has to wait for the 65nm 'Cell 2' which IBM described at Cool Chips X... I wasn't there, I've only got the one-paragraph description from the program, but the major features are that double-precision processing is now pipelined so you get 100GFLOP/chip (two flops per fmul instruction * 3.2GHz * 8 SPEs * 2 doubles per vector-register) and that the memory system is now four-channel DDR2 of up to 16GB, rather than RDRAM of up to 2GB.
[but this chip uses 100 watts, and a 16GB DDR2 memory system would be 64 chips, so it wouldn't fit in the racking or remotely in the cooling of current bluegene]
Just out of curiosity, where is the forum for interesting things done using Linux-on-PS3? I am expecting truly wonderful demos to crop up at scene.org in the medium-term future.
That's with 2.66GHz quad-cores, and it's possible that the 3GHz ones use up to 25 watts more each, but 500 watts is still a pretty pathetic space heater.
A test with 3GHz dual-cores of a server-like machine (http://www.anandtech.com/showdoc.aspx?i=2892&p=10 ) used 325 watts peak; the nVidia graphics chips do not seem to be as active at shutting down unused parts of the chip as Intel's processors are, and I think that explains most of the difference.
Yes, of course it is. But this material is not, as far as I can tell, something that you can grow in beautiful crystals for jewellery use; any applications would be as industrial superabrasive, and the fact it's ten times the price of industrial diamond dust is an issue.
I think you might be getting your pricing from one of the several Web sites that confuses rhenium (element 75) and rhodium (element 45).
Platinum is roughly $1300 per troy ounce of 31 grams on the spot market; rhenium is $13 per gram from ebay, so about $400 per troy ounce. Rhodium is $6300 per troy ounce, and so hopelessly rare and hard to extract that people have made reasonable economic arguments for reprocessing it from used reactor fuel since it's a common fission product and has no very-long-lived radio-isotopes (though you would need to leave it twenty or thirty years for the Rh-101, hl=3.3 years, to decay, which would impose nasty cost-of-capital problems).
Rhenium costs £6.50 per gram if you want to buy it on ebay; boron is £13.50 a gram on ebay because the one seller there is selling an exotic crystalline form. [ebay search for 'rhenium metal' or 'boron element']
So making ReB2 using source materials bought in small quantities on ebay would be about ten pounds (about twenty dollars) a gram; probably the cost of the electricity to run the furnace would be more than that, and the depreciation on the furnace more still.
I paid ten Euros (about fifteen dollars) for the diamond sample I have, which is two milligrams, and various diamond-industry sites give prices on the order of a hundred thousand dollars per gram; of course, rather like microchips, diamond pricing is exponential in the size because you have to find one big diamond rather than gluing two small ones together.
http://www.metalprices.com/FreeSite/metals/re/re.a sp suggests that bulk rhenium is $3000 per pound, which is a bit over half the ebay price above; some sites, I think mostly run by gold bugs, suggest $6000 per troy ounce, so either there's an opportunity for arbitrage, or they've confused rhenium and rhodium.
The not-so-trustworthy-looking http://biotsavart.tripod.com/bmt.htm has boron at about $5000 per kilogram, so $2200 per pound; still these are orders of magnitude cheaper than diamond.
as well as countermeasures; randomising the low-order bit of all your pixels anew in every frame would be ideal, but using colours which have the same number of bit transitions in 'black' and 'white' works almost as well. Looks a bit ugly to have your screen entirely in off-greens and off-pinks, but that's the price of security.
HDCP actually helps against this kind of thing, because there are no long lengths of wire carrying unencoded video signal.
Of course this new processor won't help you read faster, think faster or type faster; and e-mail, web browsing, word processing and slashdot posting have been constrained by human fingers and brains rather than by inadequate computation since the Pentium. If you're using gnome, put a System Monitor and a Frequency Monitor in your panel and see just how rare it is that the Frequency increases above bare-minimum or the load average bar leaves the very bottom of the window.
You can already open, on a five-year-old computer if you want to use current Firefox and a ten-year-old one if you're OK with Internet Explorer 3, windows enough that your short-term memory and your screen space is the limiting factor.
Faster computers mean that you can write experimental code which runs in acceptable time while concentrating on the problem domain rather than on optimisation. They enable silly hobbies - I factorise 130-digit numbers in my idle cycles, each one takes a week on a Core2Duo while in 1997 they took six months of work on a distributed system followed by three days on a Cray.
www.anandtech.com has a presumably very similar review (since these are lists of benchmarks which the journalists observed being run by Intel on Intel-provided systems), and enough bandwidth that you can actually get through to it.
It's a little annoying that these chips require different voltage regulators from the ones on current motherboards, since the chipsets are the same and changing the motherboard adds £80, some hours of fuss and an inordinate number of screws to what should be a trivial CPU upgrade, whilst bare motherboards, and even motherboard+CPU pairs, don't seem to sell well on ebay.
Slashdot Mirror
You can't patent the effect, but a USPTO search gives 77 patents on an assortment of devices (including one suggesting that you should sputter radon atoms into the disc surface - holy radioactive storage, Batman!) exploiting it, and US patent 6441661 (assigned to Fujitsu) looks as if it's on GMR magnetic sensors in general.
Does anyone have tools for traversing the graph of patents under reference in both directions? Key patents would tend to show up at the top of lists sorted by number of citations, but I don't think that's something I can sort on in the USPTO database.
That the words 'texture', 'instruction pointer' or 'blitter' appear nowhere in either PDF file is a bit of a giveaway.
Whilst the registers are essential for getting any kind of driver to work, the documents don't describe the exciting features of the graphics processor. They give you enough control over the memory-controller timings to convert any Radeon card into a smoking brick with a small kernel-mode driver, but they don't give instructions which actually make the graphics silicon do things. There's no indication of what the machine-code for the vector processors looks like.
If you compare this to the documentation that Intel has for its (obsolete) 845 graphics controller, you notice that the whole block of registers for controlling even something as basic as the blitter, let alone the 'set instruction pointer for processing unit N' registers which actually let you set the high-performance processing units in the card to work, are missing.
These documents let you use an R500 or R600 card as a frame buffer. Not worth making a song and dance about that one.
Myself, I'd be fascinated to see documentation for the Intel G965 like the documentation for the G845; it clearly exists, there's a paper in the most recent Intel Technical Journal about low-level programming on the 965, it's just not available to mortals unless by attempting to reverse-engineer the x.org 965 driver.
This was just a scheduling problem; to figure out the orbit of Dysnomia, you need to take several exposures at different times using a telescope facility capable of resolving Eris and Dysnomia. Which means either the Keck laser guide-star system, or Hubble. Each of which has a queue of astronomers reaching three times round the block who want to use them; and the queueing systems are designed for 'I want to spend a night taking spectra of this radio-galaxy' rather than 'I need pictures which take a five minute exposure, but I'd like one every 30 hours for a month'.
Also, it takes ages to get papers written and published; the Hubble pictures of Dysnomia were from August 2006 and the paper (in Science, a journal which is really quite quick to publish) came out yesterday.
The 'special crystal which took three years to grow in Russia' that the article writes about is made of pure silicon-28 precisely to get around this objection.
I think the concern is that samples of silicon from different sources (consider, for example, 'depleted silicon' from the scrapyard of the Russian isotope-enrichment facility) might have different isotope distributions at the 10^-7 level, whilst good laser enrichment can ensure a really very constant isotope distribution.
For larger-scale effects, the Miravi site has some very nice data, though I'll admit it's not as well-presented as Google Earth. It's from the MERIS instrument on the European Space Agency's Envisat, which takes a constant thousand-kilometre-wide swathe at 250 metres per pixel as it heads round the Earth in a 101-minute orbit, retracing its steps exactly every 35 days.
You can watch reservoirs filling, watch rainforest clearance in Brazil (though not so well in Borneo, since the country is almost perpetually clouded over), look at algal blooms in the North Atlantic, and see the smoke from volcanic eruptions and big fires. Over a few years you might be able to watch cities grow in China, though it's surprising how well cities blend into the landscape at 250 metres per pixel.
The data's available as JPEG files, and the only slight problem is that Firefox regards JPEG files of more than 2^15 rows as corrupt.
Romania is probably the hardest country of Eastern Europe to intimidate by means of gas supplies; it has quite substantial local production of oil (Ploesti used to be the oil capital of Europe) and of natural gas, a couple of modern nuclear reactors at Cernavoda on the Black Sea coast, and exports electricity.
Central Romania feels very energy-poor, but that's an infrastructure rather than an availability issue; it's a big place, and not a wealthy one, and they haven't yet got round to putting in the wires and the pipes universally.
Large memory modules mean otherwise-unavailable large memory capacity in a single system, since systems have only a fixed number of memory slots; top-end CPUs give you single-threaded performance which is otherwise unavailable.
But very few large-data-storage applications are heavily constrained by the number of drives in the system, so almost always you can use two small drives instead of one drive of twice the size; yes, more drives take up more space, but very rarely is the space taken up by one drive going to cost more than twice the price of the disc drive.
I could see '2.5" drive of unprecedented capacity' as a useful provision of an otherwise-unavailable facility, since laptops tend to have exactly one drive. But there are remarkably few systems that are constrained to have only one 3.5" drive.
With the ready availability of RAID cards and drivers, I wonder what the point of a terabyte drive at well above the current best cents-per-gigabyte pricing is.
This costs three times as much as a 500G drive for twice the capacity; all the reviews are pointing out that you'd want to run two drives in RAID1 because 900G of data loss from a single mechanical failure is unlikely to be acceptable; six 500G drives in RAID6 get you twice the capacity, more reliably, for the same price, three 500G drives in RAID5 get you the same capacity, similar reliability, half the price.
Yes, six drives are a little louder that two drives, but drives have been quieter than CPU/PSU fans for a long time now. I suppose some cases don't fit six drives very conveniently, but you can get quite a fancy case for $400.
I suppose there might be hosting situations in which the volume of two hard drives is worth $400.
Yes, you can make various hot-ices on Earth using a diamond-anvil cell - the pressures required tend to be in the gigapascals, so tens of thousands of atmospheres. At fairly high pressure you get ice-7, where the oxygen atoms in the water form two interpenetrating cubic lattices; at higher pressure you get ice-10 where the oxygen atoms are all lined up in a single cubic lattice.
The whole point of this patent is that you can't take the drive out and copy it, because you can't send ATA commands to the drive without authenticating yourself, and you can't authenticate yourself unless you demonstrate your ability to compute (reading between the lines of the patent) SHA1(challenge ^ magic_number_stored_in_Tivo_firmware).
Like all of this kind of authentication processes, I expect this one to last until someone gets fed up, reads off the drive firmware from the ROM chip on the board, and patches it not to do the checks. Though if Tivo have actually convinced WD to take the enormous debuggability hit from embedding the ROM in the drive-controller ASIC rather than having it as an externally-accessable device on the board, this is not a bad security approach.
That is a dreadful patent, and it would be ridiculous to see it issued; hardware challenge-response dates back to at least the first IFF machines in the second world war, they're not even mentioning having a deliberately slow password-hashing algorithm, which is itself at least as old as UNIX, and the technique is vulnerable to bump-in-the-ATA-cable extraction of the data from the disc in the first place, and probably also to an attack where you swap the drive controller board for one from a drive of similar model without Special Tivo Sauce.
I can't be the only person whose reaction to the article is to google 'David Mullins', and discover that it's a reasonably common name, shared by the professor of housing policy at Birmingham University, the director of academic administration at Warwick, a 1991 Stanford math grad student, a London-based artist, and the ex-vice-chairman of the Federal Reserve Bank of Minneapolis.
I think it's distinctly unprofessional for an organisation to record the fact that it fired someone in a document on the Google-accessible web, since having the fact available without the explanatory information that you'd get from asking the organisation for a reference might well be prejudicial to future employment for the person concerned.
NGIA is the map-making part of what used to be the National Reconnaissance Organisation, the consumers of vast amounts of black-budget money.
5 20642,47.844225&spn=0.005093,0.00824&z=17 ) - I don't know whether the bridge is still a pontoon-bridge.
i a/basra_1990.jpg
5 494,47.669935&spn=0.040728,0.065918&z=14
4 4&spn=0.007239,0.016479&t=k&z=16&om=1
There are all sorts of censorships in Google Earth, from the glaring (there are no roads or cities in Israel!) to the moderately glaring (you can count planes at Beirut airport, not so at Ben Gurion) to the subtle; I had a friend out at the RAF base in Basra a few months ago, and was a little alarmed that there was one-metre georeferenced imagery of the camp available - though since the camp's still there and the impacts on it seemed fairly random, it would appear that the local insurgents didn't have GPS-guided mortar rounds. Or possibly GPS was jammed over the camp -- after all, the airmen in the camp know where it is already -- though maintaining the navigation hardware in planes when you can't test it in-base would not be fun.
That made the British news, and maps.google.co.uk now has no GIS information for Iraq at all (in fact, no GIS from the Egyptian border to the Indian border), though it still has sub-metre imagery of central Basra
( http://maps.google.co.uk/?ie=UTF8&t=h&om=1&ll=30.
Basra International Airport, which I was only able to find after doing a Google search and finding a Soviet map of the area
http://www.lib.utexas.edu/maps/middle_east_and_as
now shows suspiciously devoid of planes and buildings:
http://maps.google.co.uk/?ie=UTF8&t=h&om=1&ll=30.
My favorite Google Earth oddity is the Mondrianised patch of the Netherlands here
http://maps.google.co.uk/?ie=UTF8&ll=52.248443,4.
Noordwijk is the home to ESA Mission Control, but that shows up on the map without trouble. I'm by a long way not interested enough to spend a hundred Euros getting myself to Holland, walking down Albert Verweystraat taking photos of the buildings, and seeing which one belies the fact that it's labelled National Marine Conservancy by the platoon of marines outside.
Don't buy 60-watt-equivalent bulbs, then. If you're concerned about brightness, buy the brightest bulbs available, they'll still use way less than 60 watts of power, and if the peak light output is 150-watt-equivalent then the brightness during the warm-up period is still usually more than the 60-watt bulb you started with.
OK, I suffer somewhat from SAD in the winter, so opt to have as much light around as possible.
Certainly there are urban legends around which are propagated by conservative propaganda sites.
... with the low power consumption, you can use cheapest-available desk lamps to put the bulbs in, and place them wherever's convenient.
But that there are five milligrams of Hg in a compact-fluorescent lightbulb is not one of them; in particular, the link that you provided admits that.
I too have a house full of CFLs - people complaining that 60-watt-equivalent CFLs are too dim are taking slightly the wrong approach, CFLs are so much more efficient than incandescent lights that you can put, into a fitting that can only handle 60 watts of heat, a 23-watt CFL which is equivalent to a 150-watt incandescent. My study is lit with three 23-watt CFLs, which provides a really excellent reading light
The permitted mercury content of tuna in Canada is 0.5 part per million, so a 170-gram tin contains at most 85 micrograms of the stuff, about a factor sixty less than the lightbulb.
I think this is more a story about how good we are at detecting minuscule quantities of material, and how political requirements tend to be of the form that the allowable amount of a dangerous material should be a small multiple of the detection limit; I would wager that the health damage caused by the stress of being told to find two thousand dollars to decontaminate your living room is significantly greater than any that could possibly be caused by five milligrams of mercury vapour.
This is because part of the point of the article is that this new equipment reduces enormously the corona-discharge losses which are producing the amazing sparks ... marvelling at the sparks in that picture is similar to marvelling at how loud your car engine is; every joule spent in sparking is one that's not being put into creating the fusion environment.
This isn't where I would have expected IBM to put Cells; from the first announcements several years ago of Cells and of the Blue Gene architecture, everyone's asked 'when do we get a Blue Gene made of Cells?'
... I wasn't there, I've only got the one-paragraph description from the program, but the major features are that double-precision processing is now pipelined so you get 100GFLOP/chip (two flops per fmul instruction * 3.2GHz * 8 SPEs * 2 doubles per vector-register) and that the memory system is now four-channel DDR2 of up to 16GB, rather than RDRAM of up to 2GB.
That may be something that has to wait for the 65nm 'Cell 2' which IBM described at Cool Chips X
[but this chip uses 100 watts, and a 16GB DDR2 memory system would be 64 chips, so it wouldn't fit in the racking or remotely in the cooling of current bluegene]
Just out of curiosity, where is the forum for interesting things done using Linux-on-PS3? I am expecting truly wonderful demos to crop up at scene.org in the medium-term future.
http://www.gamepc.com/labs/view_content.asp?id=x53 55&page=3&cookie_test=1 says 450 watts peak power consumption, for a system with two quad-core processors and a crazy nVidia graphics card.
0 ) used 325 watts peak; the nVidia graphics chips do not seem to be as active at shutting down unused parts of the chip as Intel's processors are, and I think that explains most of the difference.
That's with 2.66GHz quad-cores, and it's possible that the 3GHz ones use up to 25 watts more each, but 500 watts is still a pretty pathetic space heater.
A test with 3GHz dual-cores of a server-like machine (http://www.anandtech.com/showdoc.aspx?i=2892&p=1
Yes, of course it is. But this material is not, as far as I can tell, something that you can grow in beautiful crystals for jewellery use; any applications would be as industrial superabrasive, and the fact it's ten times the price of industrial diamond dust is an issue.
I think you might be getting your pricing from one of the several Web sites that confuses rhenium (element 75) and rhodium (element 45).
Platinum is roughly $1300 per troy ounce of 31 grams on the spot market; rhenium is $13 per gram from ebay, so about $400 per troy ounce. Rhodium is $6300 per troy ounce, and so hopelessly rare and hard to extract that people have made reasonable economic arguments for reprocessing it from used reactor fuel since it's a common fission product and has no very-long-lived radio-isotopes (though you would need to leave it twenty or thirty years for the Rh-101, hl=3.3 years, to decay, which would impose nasty cost-of-capital problems).
Rhenium costs £6.50 per gram if you want to buy it on ebay; boron is £13.50 a gram on ebay because the one seller there is selling an exotic crystalline form. [ebay search for 'rhenium metal' or 'boron element']
a mond_powder_price_list.html will sell you twenty grams (a hundred carats) of half-micron diamond dust for fifty dollars which is a lot cheaper than either the rhenium or the boron.
a sp suggests that bulk rhenium is $3000 per pound, which is a bit over half the ebay price above; some sites, I think mostly run by gold bugs, suggest $6000 per troy ounce, so either there's an opportunity for arbitrage, or they've confused rhenium and rhodium.
So making ReB2 using source materials bought in small quantities on ebay would be about ten pounds (about twenty dollars) a gram; probably the cost of the electricity to run the furnace would be more than that, and the depreciation on the furnace more still.
I paid ten Euros (about fifteen dollars) for the diamond sample I have, which is two milligrams, and various diamond-industry sites give prices on the order of a hundred thousand dollars per gram; of course, rather like microchips, diamond pricing is exponential in the size because you have to find one big diamond rather than gluing two small ones together.
But ReB2 will be competing with diamond abrasive, and http://www.diamondtech.com/products/categories/di
http://www.metalprices.com/FreeSite/metals/re/re.
The not-so-trustworthy-looking http://biotsavart.tripod.com/bmt.htm has boron at about $5000 per kilogram, so $2200 per pound; still these are orders of magnitude cheaper than diamond.
This really isn't new news; the work was done in 2004 and presented as
http://www.cl.cam.ac.uk/~mgk25/pet2004-fpd.pdf
as well as countermeasures; randomising the low-order bit of all your pixels anew in every frame would be ideal, but using colours which have the same number of bit transitions in 'black' and 'white' works almost as well. Looks a bit ugly to have your screen entirely in off-greens and off-pinks, but that's the price of security.
HDCP actually helps against this kind of thing, because there are no long lengths of wire carrying unencoded video signal.
Of course this new processor won't help you read faster, think faster or type faster; and e-mail, web browsing, word processing and slashdot posting have been constrained by human fingers and brains rather than by inadequate computation since the Pentium. If you're using gnome, put a System Monitor and a Frequency Monitor in your panel and see just how rare it is that the Frequency increases above bare-minimum or the load average bar leaves the very bottom of the window.
You can already open, on a five-year-old computer if you want to use current Firefox and a ten-year-old one if you're OK with Internet Explorer 3, windows enough that your short-term memory and your screen space is the limiting factor.
Faster computers mean that you can write experimental code which runs in acceptable time while concentrating on the problem domain rather than on optimisation. They enable silly hobbies - I factorise 130-digit numbers in my idle cycles, each one takes a week on a Core2Duo while in 1997 they took six months of work on a distributed system followed by three days on a Cray.
www.anandtech.com has a presumably very similar review (since these are lists of benchmarks which the journalists observed being run by Intel on Intel-provided systems), and enough bandwidth that you can actually get through to it.
It's a little annoying that these chips require different voltage regulators from the ones on current motherboards, since the chipsets are the same and changing the motherboard adds £80, some hours of fuss and an inordinate number of screws to what should be a trivial CPU upgrade, whilst bare motherboards, and even motherboard+CPU pairs, don't seem to sell well on ebay.