The best part of the apple disk was the dancing guy. By Bishop, IIRC.
It played some little ditty (I belive it is the "turkey jig" or some silly name) and had some guy dancing on the screen (simple page flipping, no doubt) in tune to the music, but it was molto impressive at the time.
That is a great point. I'd mod you up, but I notice that you're at 5 already, and I don;t have any mod points anyway. But here you go, you receive the smart post of the day award. Wear it with pride.
The first I ever heard about this was warez.blackdown.org (net?), which I read about on the blackdown site. 'twas an IRC log, with some semi-unqualified sysadmin threatening legal action if they didn't stop pointing that adress to "his" machines.
eventually he clued on, but not until he made a severe jackass of himself.
A google search for warez.blackdown.org should be good for a laugh, but I'm lazy.
Fess up! what did they have on you. I'm not expecting you to give details, but was it accurate, scarily precise, or just vague and uninformative.
Me, I'm proud of the fact that the credit agencies do a knock-kncok routine ("Johan who?") when they hear my name. Tho it does make getting appartments a hassle.
I've been thinking about hacking fileutils to add a switch to mv and rm to allow wiping of files before deletion. Now (in these comments) I read about fileutils 4.x including a wiper.
I dunno. shredding? I've always (*) thought that shredded documents were about as secure as... well something not too secure at all.
All you have to do is make some device to scan all the document fragments (an intern would be a cheap option) and then use std puzzle laying algorithms to reassemble the pieces. If you maintain the correspondence between front and back sides of strips, you are doing even better, as ambiguities form one side can be resolved by the other.
This should be O(n^2), I think, so it should be feasable for large inputs. Sorting the strips by "depth in bin" should speed it up even more.
Of course, this may not apply to substrates, but I suspect that shredding is still used as secure disposal of many documents.
Nah, that's not a showstopper. After all, laptops have been getting 2 to 9 hours of autonomous use and people are happy with it. All you have to do is provide screen covers and tell the implementors to not use too much blue in their graphics. That's not a problem.
Alternatively, you might sell it cheap enough to let people replace them often -- replacing your screen could become just as run-of-the-mill as replacing or charging a battery.
It's just a marketing issue. Eventually, the technology will have caught up and the replacements become less and less frequent.
Nuclear simulation, like fluid dynamics, is basically a cellular simulation -- make several bazillion cells, time step each one, communicating only with the neighbors on each step.
(I think. I'm actually making this up as I go along, so add salt to taste)
Now the problem is that since the entire simulation goes in lock-step, you limit the number of steps by not only computation speed, but also communication speed between the nodes.
I presume that there are smart approximative approaches that can be used to assauge this, but it remains the case that distribution and cellular simulation just don't go.
Plus, you eliminate the advantages of any optimization that the Code Morphing layer delivers.
and dynamo from HP showed us that these gains are not inconsiderable. Infact, it is likely the case that native/sub-morphing code would run slower than interpreted/morphed code.
I think you'll find that 90% of the processor's power is spent in busywait loops waiting for you to do something. That number is probably pretty low. This is why LongRun is such a neat idea.
You just don't need the cycles, 95% of the time. And when you do need them, the chip switches gears and runs faster.
Why are the G4s giving motorola such a hard time getting them to run fast? Complicated instruction dependencies? Not deep enough a pipeline?
If the Transmeta people are able to claw back some performance by having a simpler core run fast and then dynamically compiling it, that would be cool.
On a related note, since the TM is a VLIW proc, it should (?) have a decent chance of dealing ok with the AltiVec instructions, no?
I haven't had a look at the press release, only the AP article. The one place this would be useful is in the 2nd or 3rd level cache. If you can compress fast enough then the extra size of the cache would come in handy.
The reason why this technique would work on a cache but I can't see how it would work on main memory is that you don't ever see the cache, so you never make assumptions about how large it will be. So the encrypted data sitting in the cache fills it up after 1 Meg, while the empty matrix full of zero can have almost 4 Megs cached.
Cache is transparent, so you can never be bitten by it changing logical size on you. Unlike main memory.
For main memory, like sig11 points out, the OS will make assumptions that it can store exactly that much info. And since the hardware has no provisions for asking the OS to page out memory (and has no business asking, either) eventually havock will be wrought.
So I'm strongly suspecting that this will turn out to be used for 3rd or 4th level cache.
well, 1800 on one T1 might be really slow, 3600 on 2 might be slow too but less so, but 36000 on twenty t1s is gonna be almost useable. This is because most users will be doing very bursty things, so even though they'll want quite a bit of peak capacity -- so that the page loads quickly -- they really don't want that much bandwidth on average during the day.
It would actually be interesting to see how much external bandwidth an average person uses during a working day. Does anyone have any info on this?
Of course, I'm excluding people like myself who like to run X from home, or run any sort of trafic analysis countermeasures (cryptonuts, mostly) or who routinely suck in the whole kernel instead of the diffs... So I guess that's pretty much everyone but my mom.
yeah! how dare those pesky liberals try to pass on the community (hidden? I forget the economic term for this sort of thing) costs to the companies that benefit from it.
We need a corporate pollution tax. This is standard economics. What happens is that companies have to balance costs in order to provide competive products. So balance cost of labour against cost of more efficient manufacturing equipment. Most of these goods are priced by the market, just like the goods the company produces, so everyone wins. The environment is not, so it falls to the government to set the price of pollution.
If the government doesn't set a high enough price for pollution, the correct thing for companies to do is to switch to dirtier, and cheaper production processes.
So with that in mind, environment mania is nothing more than an attempt to correctly price an invaluable resource.
From what I gathered from a quick read overview, pretty much all of the strategies are dynamically programmable, whcih of course will cut down the time needed per turn dramatically.
D'yall remember that slashdot article about a year ago where these guys were selling vinyls with nothing but timing information on them. You hooked up the turntables to the soundcard, which then scratched mp3s (or whatever) controlled by how you were scratching the vinyl. No more lugging crateloads of discs to parties...
Slashdot Mirror
The best part of the apple disk was the dancing guy. By Bishop, IIRC.
It played some little ditty (I belive it is the "turkey jig" or some silly name) and had some guy dancing on the screen (simple page flipping, no doubt) in tune to the music, but it was molto impressive at the time.
That is a great point. I'd mod you up, but I notice that you're at 5 already, and I don;t have any mod points anyway. But here you go, you receive the smart post of the day award. Wear it with pride.
The first I ever heard about this was warez.blackdown.org (net?), which I read about on the blackdown site. 'twas an IRC log, with some semi-unqualified sysadmin threatening legal action if they didn't stop pointing that adress to "his" machines.
eventually he clued on, but not until he made a severe jackass of himself.
A google search for warez.blackdown.org should be good for a laugh, but I'm lazy.
Soo...
Fess up! what did they have on you. I'm not expecting you to give details, but was it accurate, scarily precise, or just vague and uninformative.
Me, I'm proud of the fact that the credit agencies do a knock-kncok routine ("Johan who?") when they hear my name. Tho it does make getting appartments a hassle.
I've been thinking about hacking fileutils to add a switch to mv and rm to allow wiping of files before deletion. Now (in these comments) I read about fileutils 4.x including a wiper.
Anyone know wheter the switches are there?
I dunno. shredding? I've always (*) thought that shredded documents were about as secure as... well something not too secure at all.
All you have to do is make some device to scan all the document fragments (an intern would be a cheap option) and then use std puzzle laying algorithms to reassemble the pieces. If you maintain the correspondence between front and back sides of strips, you are doing even better, as ambiguities form one side can be resolved by the other.
This should be O(n^2), I think, so it should be feasable for large inputs. Sorting the strips by "depth in bin" should speed it up even more.
Of course, this may not apply to substrates, but I suspect that shredding is still used as secure disposal of many documents.
Johan
I've read that frozen, they can retain their info for "days". This of course implies that unfrozen, they should decay more quickly.
Well, I guess you could just use encraption technologies.
well the pongo stick is why perdita had so many dalmatians.
On a slightly less feasetious note, I do recall seeing a gas powered pogo stick.
terry pratchet? One of the first ones (color of magic, light fantastic?)
18" models of Stonehenge on stage
Oh, that would be so cute! but you gotta wonder who would see them.
:-P
Nah, that's not a showstopper. After all, laptops have been getting 2 to 9 hours of autonomous use and people are happy with it. All you have to do is provide screen covers and tell the implementors to not use too much blue in their graphics. That's not a problem.
Alternatively, you might sell it cheap enough to let people replace them often -- replacing your screen could become just as run-of-the-mill as replacing or charging a battery.
It's just a marketing issue. Eventually, the technology will have caught up and the replacements become less and less frequent.
running sshd in user mode is trivial -- all you need is some sort of shell on the remote host.
The only difference from an official sshd install is that it will run on a 1024+ port and only work for you.
Johan
Nuclear simulation, like fluid dynamics, is basically a cellular simulation -- make several bazillion cells, time step each one, communicating only with the neighbors on each step.
(I think. I'm actually making this up as I go along, so add salt to taste)
Now the problem is that since the entire simulation goes in lock-step, you limit the number of steps by not only computation speed, but also communication speed between the nodes.
I presume that there are smart approximative approaches that can be used to assauge this, but it remains the case that distribution and cellular simulation just don't go.
Johan
Isn't the idea of these aggregate computers (clusters, whatever) that you can just keep on growing them?
Add a couple hundred nodes, buy another switch or two, increase your flops by a couple of hundred g?
Johan
and dynamo from HP showed us that these gains are not inconsiderable. Infact, it is likely the case that native/sub-morphing code would run slower than interpreted/morphed code.
Cute!
I think you'll find that 90% of the processor's power is spent in busywait loops waiting for you to do something. That number is probably pretty low. This is why LongRun is such a neat idea.
You just don't need the cycles, 95% of the time. And when you do need them, the chip switches gears and runs faster.
Johan
Ok, this is actually a really interesting point.
Why are the G4s giving motorola such a hard time getting them to run fast? Complicated instruction dependencies? Not deep enough a pipeline?
If the Transmeta people are able to claw back some performance by having a simpler core run fast and then dynamically compiling it, that would be cool.
On a related note, since the TM is a VLIW proc, it should (?) have a decent chance of dealing ok with the AltiVec instructions, no?
Johan
You of course know that google now has an index too? Tho I suspect it is machine classification in action.
yeah, what he said!
I haven't had a look at the press release, only the AP article. The one place this would be useful is in the 2nd or 3rd level cache. If you can compress fast enough then the extra size of the cache would come in handy.
The reason why this technique would work on a cache but I can't see how it would work on main memory is that you don't ever see the cache, so you never make assumptions about how large it will be. So the encrypted data sitting in the cache fills it up after 1 Meg, while the empty matrix full of zero can have almost 4 Megs cached.
Cache is transparent, so you can never be bitten by it changing logical size on you. Unlike main memory.
For main memory, like sig11 points out, the OS will make assumptions that it can store exactly that much info. And since the hardware has no provisions for asking the OS to page out memory (and has no business asking, either) eventually havock will be wrought.
So I'm strongly suspecting that this will turn out to be used for 3rd or 4th level cache.
:-(
I thought it was a nifty idea.
but 1.2 ms is too much !?! How do you even measure that? 1.2ms is what, 800 bps?
well, 1800 on one T1 might be really slow, 3600 on 2 might be slow too but less so, but 36000 on twenty t1s is gonna be almost useable. This is because most users will be doing very bursty things, so even though they'll want quite a bit of peak capacity -- so that the page loads quickly -- they really don't want that much bandwidth on average during the day.
It would actually be interesting to see how much external bandwidth an average person uses during a working day. Does anyone have any info on this?
Of course, I'm excluding people like myself who like to run X from home, or run any sort of trafic analysis countermeasures (cryptonuts, mostly) or who routinely suck in the whole kernel instead of the diffs... So I guess that's pretty much everyone but my mom.
yeah! how dare those pesky liberals try to pass on the community (hidden? I forget the economic term for this sort of thing) costs to the companies that benefit from it.
We need a corporate pollution tax. This is standard economics. What happens is that companies have to balance costs in order to provide competive products. So balance cost of labour against cost of more efficient manufacturing equipment. Most of these goods are priced by the market, just like the goods the company produces, so everyone wins. The environment is not, so it falls to the government to set the price of pollution.
If the government doesn't set a high enough price for pollution, the correct thing for companies to do is to switch to dirtier, and cheaper production processes.
So with that in mind, environment mania is nothing more than an attempt to correctly price an invaluable resource.
Johan
From what I gathered from a quick read overview, pretty much all of the strategies are dynamically programmable, whcih of course will cut down the time needed per turn dramatically.
Johan
2ms might be too much for monitoring, but...
it kicks ass for scratching.
D'yall remember that slashdot article about a year ago where these guys were selling vinyls with nothing but timing information on them. You hooked up the turntables to the soundcard, which then scratched mp3s (or whatever) controlled by how you were scratching the vinyl. No more lugging crateloads of discs to parties...
Now that's what you need low latency for.
Does anyone know what happened to them?