Zalman's a good brand, and I've used some of their aftermarket heatsinks. This PSU would be considerably better than a non-name grey 300W PSU, but it doesn't meet my list of desirable qualities: >80% efficiency, fanless or 120mm fan (for low noise), modular cables (for aesthetics and clean air-flow.)
This PSU looks a bit loud to me, but I'm a silence-freak. (My home computer has a single low noise fan (120mm Nexus) and even this I run at only ~60% speed.)
There is a problem that nobody seems to be interested in producing good low capacity PSUs. If you want (say) 80% efficiency and modular cabling, most manufacturers don't have anything below 500W. (A few have 400W.)
Some good options now are Seasonic S12, Antec NeoHE, Silverstone ST30NF, Nexus NX-80x0 series.
Those 600+ W power supplies are purely for people with inferiority complexes about other aspects of their lives/bodies. Here's a discussion about how much you can run on a 300W PSU. 300W suffices for a modern high-end CPU plus high-end GPU plus a bunch of drives, when under heavy load. Even a high end system will idle at around 150W. A more sensible system is probably idling around 80W.
NOTE: All the figures above are *not* including losses in the PSU. A modern PSU should be about 7 5% efficient, so increase the above by 1/3 to make them comparable to the 400W number in the article.
Furthermore, if the software can detect the password cadence, so can an appropriately programmed keylogger.
Almost all security is a tradeoff against usability. This one looks like a bad trade - you lose lots of usability for only a small increase in security.
If they did, there is a pretty good chance of getting caught - slip-ups in their story, other survelence cameras, a police car that just happened to be nearby when the alarm went out... The odds are probably at least as good as catching a real ATM mugger.
Chimps live in nature... Humans... shape the environment This doesn't prevent/slow human evolution compared to chimp evolution - it just means that the evolutionary pressures are different. Humans gain more from being good at politics than chimps do, chimps gain more from being able to survive periods of bad weather without shelter.* (Also - humans have only been shaping the environment for a very small portion of the ~5 million years since the human/chimp ancestor.)
Here are some factors which theoretically make a difference: Population size: selection works more efficiently in larger populations, as random drift has proportionately less effect. (Pop size has no effect on 'neutral' mutations, however.) Generation time: more generations => more evolution Mutation rate: more mutations => more evolution (but too much gets you into trouble.) There may be a correlation between mutation rate and metabolic rate. Number of children: more children per individual => more evolution Predictability of environment: the greater the influence of random chance (as opposed to fitness) on who lives and breeds, the less efficient selection becomes.
The last two tend to counteract each other: chancy environments encourage a strategy of having many offspring.
In practice, the population size/generation time/metabolic rate effects are somewhat controversial, and certainly are smaller than a simple theory would predict.
In very recent times, medicine and birth control technology has significantly shifted the fitness criteria for humans. Poor eyesight, diabetes, cancer-proneness etc. have become much less important, and having lots of kids (either through choice or lack of forethought) much more important. As long as we have differential reproductive success, evolution will still be with us.
(* but chimps are still politicians** and humans still survive bad weather outdoors - this is just a relative ranking.) (** no, not human politics, I mean intra-tribe chimp politics.)
I'd have thought that flash was slow enough that this difference was insignificant - but I'm really just guessing here. Does anyone have some real numbers for write-to-flash latency vs HDD interface latency?
If the cache is write-through then you wouldn't have to worry about drives getting corrupted.
Fair point.
You can remove a hard disk from a notebook, but it's not exactly easy so I don't think that's much of a concern.
Power saving is just one reason for flash cache, so it isn't (or shouldn't be) restricted to notebooks. On desktop, the major advantage would be boot speed, with general speed and noise reduction as secondary benefits. I expect you'd use more of it for reading and less for writing than for a laptop, as you're going to be keeping the drive spinning anyhow. (For me, the noise reduction would be a major advantage, but I'm a computer silence nut.)
I think the real reason Robson exists is political: Intel doesn't make hard disks, but they want to offer some kind of flash cache, so they build what they know how to build.
There are three technologies (that I'm aware of) for using flash to cache disk. There are 'hybrid drives' where the flash is part of the hard drive, there's the Windows Vista method which uses a separately attached flash memory (typically USB), and there is this "Robson Technology" where it is on the motherboard.
It really seems to me that the 'hybrid drive' is the Right Thing to do. The cache contents is useless without the drive, and the drive is potentially corrupt without the cache contents, so why make them separable? With appropriate firmware, the hybrid drive can make the existence of the cache transparent to the OS, so no OS support is required (but you can allow the OS finer control over the cache if it does support it.) You also automatically add more cache as you add more drives.
(Incidentally, I hope MS doesn't have a patent on this - I thought of it years ago, and I'm not even an engineer.)
I can see the Windows method as a useful 'stop-gap' to get the benefit with a non-hybrid drive, but if you're buying new hardware anyway, why would you want to put the cache on the motherboard instead of the hard drive? The only advantage I can think of is that if you have multiple drives, you can dynamically allocate how much cache is associated with each drive, according to usage patterns.
As G.R. and Q.M. are contradictory, there must be a more complete theory (the "Grand Unified Theory") which simplifies to G.R. in the limit of large sizes and to Q.M. in the limit of small gravitational potentials.
We already see many other such simplifications in physics: Quantum mechanics simplifes to Newtonian mechanics in the limit of large sizes. General relativity simplifies to special relativity in the limit of no gravity. General relativity simplifies to Newtonian gravity in the limit of low velocities and low gravitational potential. Special relativity simplifies to Newtonian mechanics in the limit of low velocities.
The greatest challenge of modern physics is to find the GUT. We can't observe the conditions where it would be obvious (high gravitational potential and small size) because this is many orders of magnitude beyond the capabilities of our accelerators. One approach is to look for promising mathematical theories which simplify in the required ways to QM and GR, and then try to find predictions these theories make which we can observe. Another approach is to subject QM and GR to ever tougher tests, in the hopes of finding a discrepancy which will point the way to a theory.
I had to fix a 100% genuine Y2K bug. I was doing (among other things) source control admin for a company of perhaps ~80 developers. At Y2K, we were using SCCS for source control. (Later changed to ClearCase, but that is a different story.) I was called in on I think 2 Jan 2000. Some eager developers had returned to work early, and their new checkins were messed up.
Although we'd updated all the computers to a Y2K-compliant version of the OS (IRIX), on one of the machines the (non-Y2K) SCCS binaries had got there by copying rather than a proper install - so the OS upgrade didn't know they were there, and didn't upgrade them to the Y2K fixed versions.
End result: I edited the corrupted SCCS files to fix them, and called a sysadmin to fix the binaries. Two people called in, some developer time lost - it probably cost about 10 geek-hours in total. I think I might have got a few hundred dollars extra pay as well - I can't remember now.
This was the Next Big Thing in the early 80s, aimed at the niche currently occupied by flash memory - non-volatile, midway between RAM and disk in terms of speed and price. Companies like Texas Instruments and Intel sunk large quantities of money into R&D. Few products got to market and those that did faded very quickly. Basically hard drives just caught up in speed but at much lower price.
I do wonder if, had the technology been a few years earlier, whether it could have carved out a niche. If it were shipping in large volume, could the correspondingly high R&D spend have kept it competitive? In an alternate history, could we be using (orders of magnitude improved) bubble memory today instead of flash?
All you need do is assemble a Great Pyramid sized pile of sand, stop it from blowing away for 20 years, make it firm enough to carry huge stones on rollers and remove it afterwards.
It looks like a gimick to get people to download their 3D viewing software (but I'm not sure, as I'm on Linux/Firefox, so all I know is that the required plugin is not available for me.) At least, that is the interest of the hosting company - the architect story may be legitimate.
The planes have a certified max takeoff weight, and they takeoff with almost exactly that weight on many if not most flights.
It is a bit more complicated, in that the max takeoff weight depends on runway length, temperature, wind speed and direction, and possibly other factors as well. (I suspect you knew that, and were deliberately simplifying.)
I'm sure the 90 m^3 is the uncompressed volume, whereas I think the grandparent post is talking about a compressed volume of 300 l. 200 bar air is 200 atmospheres (close enough) so this is 300 x 200 = 60,000 l (uncompressed volume) having 35MJ of energy*. So TFA's 90 m^3 would correspond to about 50MJ of energy, or 1.5 l of petrol*. (Maybe the GP typoed 200 for 300?)
* I'm accepting the GP's energy density figures, as I'm too lazy to check them.
Look at how much stuff in the SCO case is 'under seal' and we don't get to see it. I strongly suspect that this information will not become public. (But IANAL.)
And will the developers move to Outer Elbonia too? If MS had a (e.g.) U.S. court declare that Reactos was an Unclean Abomination That None May Look Upon, I'm pretty sure that the courts would frown on a U.S. resident developer who continued to work on it and claimed US laws didn't apply because they submitted the patches to a server in Outer Elbonia.
Great, then it comes time to merge the HR branch of the bill with the Senate branch. Some overworked staffer sits there clicking 'accept' to all changes without looking at them, then they pass the bill and wait for it to crash to find the inconsistencies.
I was wondering about this. It is very hard to really cheat at most casino games. (Note that I've set foot in a casino twice in my life, and then just for a look around, so I'm not an expert.) There are things like nicking other people's chips or cheating in poker (if the casino provides for poker games.)
My bookmark to Google uses bork-bork-bork, and I didn't stop to think that this would carry over with the URL. (If I get to Google by other means, it uses my browser language preferences, and so is in Latin. (Although my Latin is very rudimentary.))
Slashdot Mirror
Bwahahahaha! Come, join us on the dark side! Your descent into madness has begun!
BTW, there is a specialist retailier in NZ: http://www.quietpc.com/nz/. Another good shop is http://www.ascent.co.nz/, as they have a huge range of computer accessories.
Zalman's a good brand, and I've used some of their aftermarket heatsinks. This PSU would be considerably better than a non-name grey 300W PSU, but it doesn't meet my list of desirable qualities: >80% efficiency, fanless or 120mm fan (for low noise), modular cables (for aesthetics and clean air-flow.)
This PSU looks a bit loud to me, but I'm a silence-freak. (My home computer has a single low noise fan (120mm Nexus) and even this I run at only ~60% speed.)
There is a problem that nobody seems to be interested in producing good low capacity PSUs. If you want (say) 80% efficiency and modular cabling, most manufacturers don't have anything below 500W. (A few have 400W.)
Some good options now are Seasonic S12, Antec NeoHE, Silverstone ST30NF, Nexus NX-80x0 series.
Those 600+ W power supplies are purely for people with inferiority complexes about other aspects of their lives/bodies. Here's a discussion about how much you can run on a 300W PSU. 300W suffices for a modern high-end CPU plus high-end GPU plus a bunch of drives, when under heavy load. Even a high end system will idle at around 150W. A more sensible system is probably idling around 80W.
NOTE: All the figures above are *not* including losses in the PSU. A modern PSU should be about 7 5% efficient, so increase the above by 1/3 to make them comparable to the 400W number in the article.
Furthermore, if the software can detect the password cadence, so can an appropriately programmed keylogger.
Almost all security is a tradeoff against usability. This one looks like a bad trade - you lose lots of usability for only a small increase in security.
Who said they got the "stolen" money refunded?
If they did, there is a pretty good chance of getting caught - slip-ups in their story, other survelence cameras, a police car that just happened to be nearby when the alarm went out... The odds are probably at least as good as catching a real ATM mugger.
Chimps live in nature ... Humans ... shape the environment
This doesn't prevent/slow human evolution compared to chimp evolution - it just means that the evolutionary pressures are different. Humans gain more from being good at politics than chimps do, chimps gain more from being able to survive periods of bad weather without shelter.* (Also - humans have only been shaping the environment for a very small portion of the ~5 million years since the human/chimp ancestor.)
Here are some factors which theoretically make a difference:
Population size: selection works more efficiently in larger populations, as random drift has proportionately less effect. (Pop size has no effect on 'neutral' mutations, however.)
Generation time: more generations => more evolution
Mutation rate: more mutations => more evolution (but too much gets you into trouble.) There may be a correlation between mutation rate and metabolic rate.
Number of children: more children per individual => more evolution
Predictability of environment: the greater the influence of random chance (as opposed to fitness) on who lives and breeds, the less efficient selection becomes.
The last two tend to counteract each other: chancy environments encourage a strategy of having many offspring.
In practice, the population size/generation time/metabolic rate effects are somewhat controversial, and certainly are smaller than a simple theory would predict.
In very recent times, medicine and birth control technology has significantly shifted the fitness criteria for humans. Poor eyesight, diabetes, cancer-proneness etc. have become much less important, and having lots of kids (either through choice or lack of forethought) much more important. As long as we have differential reproductive success, evolution will still be with us.
(* but chimps are still politicians** and humans still survive bad weather outdoors - this is just a relative ranking.)
(** no, not human politics, I mean intra-tribe chimp politics.)
I'd have thought that flash was slow enough that this difference was insignificant - but I'm really just guessing here. Does anyone have some real numbers for write-to-flash latency vs HDD interface latency?
If the cache is write-through then you wouldn't have to worry about drives getting corrupted.
Fair point.
You can remove a hard disk from a notebook, but it's not exactly easy so I don't think that's much of a concern.
Power saving is just one reason for flash cache, so it isn't (or shouldn't be) restricted to notebooks. On desktop, the major advantage would be boot speed, with general speed and noise reduction as secondary benefits. I expect you'd use more of it for reading and less for writing than for a laptop, as you're going to be keeping the drive spinning anyhow. (For me, the noise reduction would be a major advantage, but I'm a computer silence nut.)
I think the real reason Robson exists is political: Intel doesn't make hard disks, but they want to offer some kind of flash cache, so they build what they know how to build.
That is my opinion too.
There are three technologies (that I'm aware of) for using flash to cache disk. There are 'hybrid drives' where the flash is part of the hard drive, there's the Windows Vista method which uses a separately attached flash memory (typically USB), and there is this "Robson Technology" where it is on the motherboard.
It really seems to me that the 'hybrid drive' is the Right Thing to do. The cache contents is useless without the drive, and the drive is potentially corrupt without the cache contents, so why make them separable? With appropriate firmware, the hybrid drive can make the existence of the cache transparent to the OS, so no OS support is required (but you can allow the OS finer control over the cache if it does support it.) You also automatically add more cache as you add more drives.
(Incidentally, I hope MS doesn't have a patent on this - I thought of it years ago, and I'm not even an engineer.)
I can see the Windows method as a useful 'stop-gap' to get the benefit with a non-hybrid drive, but if you're buying new hardware anyway, why would you want to put the cache on the motherboard instead of the hard drive? The only advantage I can think of is that if you have multiple drives, you can dynamically allocate how much cache is associated with each drive, according to usage patterns.
As G.R. and Q.M. are contradictory, there must be a more complete theory (the "Grand Unified Theory") which simplifies to G.R. in the limit of large sizes and to Q.M. in the limit of small gravitational potentials.
We already see many other such simplifications in physics: Quantum mechanics simplifes to Newtonian mechanics in the limit of large sizes. General relativity simplifies to special relativity in the limit of no gravity. General relativity simplifies to Newtonian gravity in the limit of low velocities and low gravitational potential. Special relativity simplifies to Newtonian mechanics in the limit of low velocities.
The greatest challenge of modern physics is to find the GUT. We can't observe the conditions where it would be obvious (high gravitational potential and small size) because this is many orders of magnitude beyond the capabilities of our accelerators. One approach is to look for promising mathematical theories which simplify in the required ways to QM and GR, and then try to find predictions these theories make which we can observe. Another approach is to subject QM and GR to ever tougher tests, in the hopes of finding a discrepancy which will point the way to a theory.
I had to fix a 100% genuine Y2K bug. I was doing (among other things) source control admin for a company of perhaps ~80 developers. At Y2K, we were using SCCS for source control. (Later changed to ClearCase, but that is a different story.) I was called in on I think 2 Jan 2000. Some eager developers had returned to work early, and their new checkins were messed up.
Although we'd updated all the computers to a Y2K-compliant version of the OS (IRIX), on one of the machines the (non-Y2K) SCCS binaries had got there by copying rather than a proper install - so the OS upgrade didn't know they were there, and didn't upgrade them to the Y2K fixed versions.
End result: I edited the corrupted SCCS files to fix them, and called a sysadmin to fix the binaries. Two people called in, some developer time lost - it probably cost about 10 geek-hours in total. I think I might have got a few hundred dollars extra pay as well - I can't remember now.
This was the Next Big Thing in the early 80s, aimed at the niche currently occupied by flash memory - non-volatile, midway between RAM and disk in terms of speed and price. Companies like Texas Instruments and Intel sunk large quantities of money into R&D. Few products got to market and those that did faded very quickly. Basically hard drives just caught up in speed but at much lower price.
http://en.wikipedia.org/wiki/Bubble_memory
I do wonder if, had the technology been a few years earlier, whether it could have carved out a niche. If it were shipping in large volume, could the correspondingly high R&D spend have kept it competitive? In an alternate history, could we be using (orders of magnitude improved) bubble memory today instead of flash?
All you need do is assemble a Great Pyramid sized pile of sand, stop it from blowing away for 20 years, make it firm enough to carry huge stones on rollers and remove it afterwards.
It looks like a gimick to get people to download their 3D viewing software (but I'm not sure, as I'm on Linux/Firefox, so all I know is that the required plugin is not available for me.) At least, that is the interest of the hosting company - the architect story may be legitimate.
The planes have a certified max takeoff weight, and they takeoff with almost exactly that weight on many if not most flights.
It is a bit more complicated, in that the max takeoff weight depends on runway length, temperature, wind speed and direction, and possibly other factors as well. (I suspect you knew that, and were deliberately simplifying.)
I'm sure the 90 m^3 is the uncompressed volume, whereas I think the grandparent post is talking about a compressed volume of 300 l. 200 bar air is 200 atmospheres (close enough) so this is 300 x 200 = 60,000 l (uncompressed volume) having 35MJ of energy*. So TFA's 90 m^3 would correspond to about 50MJ of energy, or 1.5 l of petrol*. (Maybe the GP typoed 200 for 300?)
* I'm accepting the GP's energy density figures, as I'm too lazy to check them.
Alan Turing nearly qualified for the Marathon in the 1948 Olympics. One could argue that he was also the world's first sysadmin.
Will the RIAA's costs (or even the size of the defendent's eventual payout) become public?
Look at how much stuff in the SCO case is 'under seal' and we don't get to see it. I strongly suspect that this information will not become public. (But IANAL.)
And will the developers move to Outer Elbonia too? If MS had a (e.g.) U.S. court declare that Reactos was an Unclean Abomination That None May Look Upon, I'm pretty sure that the courts would frown on a U.S. resident developer who continued to work on it and claimed US laws didn't apply because they submitted the patches to a server in Outer Elbonia.
IANAL
Great, then it comes time to merge the HR branch of the bill with the Senate branch. Some overworked staffer sits there clicking 'accept' to all changes without looking at them, then they pass the bill and wait for it to crash to find the inconsistencies.
I was wondering about this. It is very hard to really cheat at most casino games. (Note that I've set foot in a casino twice in my life, and then just for a look around, so I'm not an expert.) There are things like nicking other people's chips or cheating in poker (if the casino provides for poker games.)
My bookmark to Google uses bork-bork-bork, and I didn't stop to think that this would carry over with the URL. (If I get to Google by other means, it uses my browser language preferences, and so is in Latin. (Although my Latin is very rudimentary.))
This recursion is beginning to sound like Lewis Carroll's Paradox. (He was a mathematician/logician in his day job.)