It's rediculously late me to post a reply and expect anyone to
read it, but what the heck...
As are all modern processors.
I'm sorry, but you're incorrect.
IA-32 based CPU's do not...
Sorry, I've been living in the SPARC world too long.:)
But the thing is, there's a lot you can't know about how most programs are going to behave until run-time. Much of the available parallelism *has* to be determined by the processor itself.
I don't think you understand code generation, that, or you're being naive. The compiler is in the best position to determine what code can execute in parallel, and how best to order the instructions to gain maximum speed from the CPU in executing that code....
True, I don't have a lot of experience with the details of compiler techniques. But I have to say, I think there's a certain amount of naivete in your statement. Can compilers really be so good at branch-prediction?
Data-dependent branches are one of the fundamental features of computer programs. Heck, it's practically the whole point. As such, there will always be situations where the probabilities of which way a branch will go cannot be known at compile time; only run time. As such, a trace cache or some such will achieve better instruction-level parallelism than compiler predictions in many situations.
Now, to be fair, you seem to be talking about a higher level of parallelism than simple ILP. If a compiler can eeke that out of a program, it can put the resulting parallel instruction streams side-by-side for a VLIW machine.
But of course, if it could do that, it could also interleave such instructions for a non-vliw machine to grab out of its lookahead buffer. So the two types of machines are equal in that respect, except that when the code ends up with a lot of compiler mis-predicts on branching, the VLIW machine is screwed, whereas a trace-cache or such technique on the other machine would maintain performance, after the cache adjusts to the actual situation. And a non-vliw machine could still be capable of listening to compiler hints about which way a given branch is expected to go, if it's not in the trace cache yet. (Of course, I guess the IA-32 probably wouldn't support such communication. Don't know about AMD64.)
The other half "or prevent the free excercise thereof" was blatantly violated by the 9th circus in their ruling.
You're completely backwards there. The Pledge of Allegance, teacher-led in a classroom, cannot be considered a "free" act -- it is coerced. Even though technically a student may opt out of it, peer pressure makes this impossible most of the time. As such, the 9th circut's ruling was a blow for the free exercise of relilgion.
You have to remember that the term "religion" means more than just judeo/christian/muslim beliefs. It also encompasses beleif systems in which God is thought not to exist, or at least is not to be worshipped.
Think anything with big for loops. Think signal processing. GIMP and Photoshop filters could almost certainly be reengineered to use wide parallelism. Weather prediction models would benefit as well.
If those applications were important enough to get everyone to shift to a different instruction set, wouldn't everyone have gone to PowerPC by now (think Altavec)?
The design of the IA-64 architecture is such that it scales well with larger and larger die space.
That is, die space that's not filled with cache. The enormous amount of cache on the chip is one of the main reasons for a lot of the Itanium's performance wins.
It's designed to be able to execute code in parallel
As are all modern processors. They all use a "lookahead" strategy to find instructions further along in the stream that could be executed at the same time Branching causes complications, of course, but there are a lot of techniques (e.g. trace cache) for making good predictions about which way a given branch will go.
The compiler can help out, too, by arranging the instruction stream so as to make the parallelizable operations more accessible.
The Itanium technique, from what I understand, requires the compiler to do just about everything. The compiler *has* to find the available parallelism so that it can arrange the operations side-by-side in the same ultra-wide instruction. From there, the execution is relatively simple -- the processor doesn't have to worry about finding the parallelism, the compiler's figured it all out already! So that would certainly make it easy to scale to larger available die space (i.e. just throw on more execution units). But the thing is, there's a lot you can't know about how most programs are going to behave until run-time. Much of the available parallelism *has* to be determined by the processor itself.
That's why simple techniques to add parallelism on a single die aren't so commonplace yet -- more parallelism can be gained by adding to the *intelligence* on the chip than by just throwing on more execution units, or having multiple processors on a chip, etc.
Oh JEEZUS!!!!! Tibet was a protectorate under imperial China for hundreds and hundreds of years.
Everybody wants to 'save' Tibet, but do you have any idea of the kinds of anal rape that little child 'priests' must undergo? Tibet isn't all smiles and prayer wheels, there's a serious, rape-oriented, pedophillic culture there that's actually quite sickening.
And they practiced slavery, from what I understand -- one's priest was also one's owner? Something like that?
The modding of the parent message as "Troll" was completely unfair, in my opinion.
China has long been a serious threat to world security...
I'm sorry, but that's a rather funny statement, coming from an American these days...
But anyways, since the collapse of communisum, I haven't heard of any expansionist tendencies on the part of China. In the matters of Taiwan and Tibet, China actually has pretty substantial "prior ownership" claims. Aside from that, and internal issues such as govt. corruption and a lack of civil rights, the culture of China is actually quite pacifist -- just look into Confucionism to see what I mean.
What are they doing that would threaten "world security"?
In order for your brain to properly "fuse" the images together, your eyes will have to perform some tiresome calisthenics...
Ah, no, the whole point is that you'll focus in front of the screen where the virual object will appear, as clearly depicted in figure 1. Because your eyes will be crossed, the image to be sent to your right eye will perforce be presented on the left side of the screen, and vice versa for the left eye.
You've missed something: yes, you'll be focusing in front of the screen on the virtual object, because your eyes are crossed. But the problem is, you actually need to be focusing a good deal further from you than that -- on the screen -- in order to have the image in focus. To do that, you have to overcome a lifetime of association between how much your eyes are crossed and where your eyes are focused -- and that's the "tiresome calisthenics".
Not that that's necessarily a bad thing, mind you -- once you get used to it, it's really not so tiresome. Here's a neat little image I rendered that makes use of this technique. It's also a great way to really see
faraway scenerey like mountains. For example, I take one picture of Half Dome, then walk 50 feet to the left, and take another; at home, I put the two prints side-by-side, and I can see it in 3D! It also works pretty well out airplane windows, since you're looking perpendicular to the line of travel.
But anyways, the original poster is right -- the cellophane is really not a significant step beyond the simple eye-crossing technique by itself.
... There's a 1 in 2191.78 chance that you'll be sued by
the RIAA this year (if you're a file trader).
Those odds are high enough to have the "chilling effect"
that the RIAA seeks. Even though the odds aren't *that* high,
they're easily within comprehension, and considering the severe
pain of being sued... well, I know I wouldn't want to risk it.
I actually aggree with this approach to an extent -- it's the
"carrot and stick" approach. Of course, there's the little
matter of there being NO CARROT to go with this "stick". Which
means this is going to be a much harder fight than they'd have
on their hands otherwise. They really need to wake up to reality
and play the game right.
... that there is only one Son of God and that only through Him is salvation possible. This would mean that if by some chance that there were lifeforms elsewhere in the galaxy that they would have to be perfect beings or destined for Hell.
Naa, it just means that Earth got the male; the rest were female, or other.
If that current flows through your heart, it can be more likely to give your heart fibrillations.
Reminds me of the tale told by my high-school physics teacher, of a man jogging (or some such) in a lighning storm. A tree branch or something knocks a live powerline loose, which happens to hit him, sending his heart into fibrillations. As he was lying on the ground, almost assuredly going to die, lightning happened to strike an object nearby, whacking him with a nice, constant voltage. This "rebooted" his heart, essentially, just like a defibrillator at a hospital. It started beating normally again, and the guy was ok.
As it stands now, destroying the RFID tag (assuming you can find/get to it) is the only way you can be sure that it will not continue to allow you and your purchases to be tracked.
Maybe I should RTFA, but it seems to me that the RFID tags would naturally be destroyed at the checkout counter -- that's what they do with current RF security tags (destroy or remove them). If it can't be done forcibly with an overload, then design them to respond to a "suicide" command.
If it's a concern that the chips will merely be disabled for purposes of letting them go out of the store, but still be active in other ways, I'd say a few lawsuits would cure any company of such desires. e.g. if Wal-Mart put them in shoes, all it takes is one advocate to buy a pair, cut out the RFID tag, and prove that it's still active. Why would any company risk the embarassment?
I was kind of expecting to see some nice photos of the animals in there natural habitat where they would generally look a lot better.
Yeah, me too.
But of course, then I remember the possibility that the use of floodlights that far down can also be damaging: any light-sensing organs on any creatures down there are going to be extremely sensitive, and will likely be fried by the amount of light needed to take a picture (with standard equipment).
So yes, a picture of one of these critters swimming/floating in their native habitat would be nicer than seeing them dead on an examination tray. But it's not so nice to know that the picture shows a creature that has just been permanently blinded by the very fact that you are looking at it.
I am outraged at the state government. I am more outraged at every absolute moron that RE-ELECTED Governor Davis. Worst governor in state history.
Can you name your local representative to the state assembly?
I know I can't. I don't think many can. And that's where the deeper problem lies. It's easy to blame the Governor; it's a lot harder to study who the assemblypersons are and figure out which ones (if any) really deserve our votes.
I have a filter on my kitchen sink, which helps a bit. But buying bottled water by the case instead of by the bottle makes the price not so terrible. Certainly comparable to other store-bought beverages.
Might want to try a reverse-osmosis system. A cheap one will run
not much more than $100. Have to replace the carbon filters every
6 months, at abour $40 each time, and the reverse-osmosis filter itself every couple of years; not sure how much that is. Not too hard to set up, especially
if your kitchen sink has some pre-made accessory holes.
Probably less expensive than your current approach sounds.
...
Even without reference to China's lunar ambitions, the military implications of Shenzhou should give one pause. The integration of technologies achieved by Shenzhou , including recoverable satellite capability, implies the ability to hit targets in the United States with nuclear warheads with a great deal of accuracy....
[etc. etc.]
It is fine for USA to have space technologies with many military applications, or to have the ability to hit targets around the world accurately. But if another country does this, it threatens world peace. Why?
I was an intern at a not-to-be-named company back in the early 90's. One of the employees had a party at his house on the lake. At one point, he was showing off his nifty little powerboat to me and 2 other interns. We interns were in our day clothes. There were no seat belts, only little handhold straps. And the water was a bit choppy. As you might expect, we 3 were pretty nervous as he raced the boat around, pulling sharp turns, and bouncing us around on the water. We felt like we were about to fly out of the boat!
Then, at one point, he stopped the boat, and said "ok, who wants to take the wheel now?" Having had experience with jet skiis, which are similar but smaller watercraft, I volunteered. After I started getting the hang of it, it started to be really fun, and I even started intentionally trying to do little jumps off the wakes of other boats and such. I no longer felt in danger of flying out of the boat, despite the fact that the only thing holding me in was my grip on the wheel.
After we got back to shore, one of the other interns gave his opinion: "yeah, it was pretty scary, with him bouncing us all over the place like that, but dude, you were way worse! You were *trying* to jump those waves!!"
Moral of the story: when you're in the driver's seat, you can trust your intentions, and your ability to predict the outcome of your actions. When someone else is running things, it's a lot harder not to be nervous.
The author of the original paragraph is obviously perceiving the world in this manner. His government, whose actions he is apparently willing treat as a proxy of his own, can be seen as "trusted'; others cannot. This is someone who needs to take the time to consider the other side of this symmetry.
[I thought] that the space elevator was the new way to go? Except for the massive production of the required carbon tethers it all seems feasible and why not give it a try?
> This worm appears to primarily affect Microsoft systems
<Nelson>
Ha - Haah!
</Nelson>
And now...
<Hanz&Franz>
Once again, ha haa! I lauugh at you silly foolz, with your flabby Windowz and your buuggy virus-baiiting Outlook email reader. I sit here with my puuumped-up Linux system, and my maanly Mutt text-only mail reader, and I open up my spam and virus emails and lauugh again because they cannot haarm me!
Slashdot Mirror
IA-32 based CPU's do not ...
Sorry, I've been living in the SPARC world too long. :)
I don't think you understand code generation, that, or you're being naive. The compiler is in the best position to determine what code can execute in parallel, and how best to order the instructions to gain maximum speed from the CPU in executing that code.True, I don't have a lot of experience with the details of compiler techniques. But I have to say, I think there's a certain amount of naivete in your statement. Can compilers really be so good at branch-prediction?
Data-dependent branches are one of the fundamental features of computer programs. Heck, it's practically the whole point. As such, there will always be situations where the probabilities of which way a branch will go cannot be known at compile time; only run time. As such, a trace cache or some such will achieve better instruction-level parallelism than compiler predictions in many situations.
Now, to be fair, you seem to be talking about a higher level of parallelism than simple ILP. If a compiler can eeke that out of a program, it can put the resulting parallel instruction streams side-by-side for a VLIW machine.
But of course, if it could do that, it could also interleave such instructions for a non-vliw machine to grab out of its lookahead buffer. So the two types of machines are equal in that respect, except that when the code ends up with a lot of compiler mis-predicts on branching, the VLIW machine is screwed, whereas a trace-cache or such technique on the other machine would maintain performance, after the cache adjusts to the actual situation. And a non-vliw machine could still be capable of listening to compiler hints about which way a given branch is expected to go, if it's not in the trace cache yet. (Of course, I guess the IA-32 probably wouldn't support such communication. Don't know about AMD64.)
IA-64 is the future of Intel desktop CPU's
God, I hope so. AMD needs the market share! ;)
You're completely backwards there. The Pledge of Allegance, teacher-led in a classroom, cannot be considered a "free" act -- it is coerced. Even though technically a student may opt out of it, peer pressure makes this impossible most of the time. As such, the 9th circut's ruling was a blow for the free exercise of relilgion.
You have to remember that the term "religion" means more than just judeo/christian/muslim beliefs. It also encompasses beleif systems in which God is thought not to exist, or at least is not to be worshipped.
If those applications were important enough to get everyone to shift to a different instruction set, wouldn't everyone have gone to PowerPC by now (think Altavec)?
That is, die space that's not filled with cache. The enormous amount of cache on the chip is one of the main reasons for a lot of the Itanium's performance wins.
It's designed to be able to execute code in parallel
As are all modern processors. They all use a "lookahead" strategy to find instructions further along in the stream that could be executed at the same time Branching causes complications, of course, but there are a lot of techniques (e.g. trace cache) for making good predictions about which way a given branch will go. The compiler can help out, too, by arranging the instruction stream so as to make the parallelizable operations more accessible.
The Itanium technique, from what I understand, requires the compiler to do just about everything. The compiler *has* to find the available parallelism so that it can arrange the operations side-by-side in the same ultra-wide instruction. From there, the execution is relatively simple -- the processor doesn't have to worry about finding the parallelism, the compiler's figured it all out already! So that would certainly make it easy to scale to larger available die space (i.e. just throw on more execution units). But the thing is, there's a lot you can't know about how most programs are going to behave until run-time. Much of the available parallelism *has* to be determined by the processor itself.
That's why simple techniques to add parallelism on a single die aren't so commonplace yet -- more parallelism can be gained by adding to the *intelligence* on the chip than by just throwing on more execution units, or having multiple processors on a chip, etc.
Everybody wants to 'save' Tibet, but do you have any idea of the kinds of anal rape that little child 'priests' must undergo? Tibet isn't all smiles and prayer wheels, there's a serious, rape-oriented, pedophillic culture there that's actually quite sickening.
And they practiced slavery, from what I understand -- one's priest was also one's owner? Something like that?
The modding of the parent message as "Troll" was completely unfair, in my opinion.
I'm sorry, but that's a rather funny statement, coming from an American these days...
But anyways, since the collapse of communisum, I haven't heard of any expansionist tendencies on the part of China. In the matters of Taiwan and Tibet, China actually has pretty substantial "prior ownership" claims. Aside from that, and internal issues such as govt. corruption and a lack of civil rights, the culture of China is actually quite pacifist -- just look into Confucionism to see what I mean.
What are they doing that would threaten "world security"?
You've missed something: yes, you'll be focusing in front of the screen on the virtual object, because your eyes are crossed. But the problem is, you actually need to be focusing a good deal further from you than that -- on the screen -- in order to have the image in focus. To do that, you have to overcome a lifetime of association between how much your eyes are crossed and where your eyes are focused -- and that's the "tiresome calisthenics".
Not that that's necessarily a bad thing, mind you -- once you get used to it, it's really not so tiresome. Here's a neat little image I rendered that makes use of this technique. It's also a great way to really see faraway scenerey like mountains. For example, I take one picture of Half Dome, then walk 50 feet to the left, and take another; at home, I put the two prints side-by-side, and I can see it in 3D! It also works pretty well out airplane windows, since you're looking perpendicular to the line of travel.
But anyways, the original poster is right -- the cellophane is really not a significant step beyond the simple eye-crossing technique by itself.
Those odds are high enough to have the "chilling effect" that the RIAA seeks. Even though the odds aren't *that* high, they're easily within comprehension, and considering the severe pain of being sued... well, I know I wouldn't want to risk it.
I actually aggree with this approach to an extent -- it's the "carrot and stick" approach. Of course, there's the little matter of there being NO CARROT to go with this "stick". Which means this is going to be a much harder fight than they'd have on their hands otherwise. They really need to wake up to reality and play the game right.
Or would that be "heteroamohomosexualist"?
Hmmm... naa, sometimes a joke works better in its implied form.
Except when the fighters launch. :)
Naa, it just means that Earth got the male; the rest were female, or other.
I would think that someone with the proper equipment would be able to elicit a response from it, and register the radio waves it sends out.
Reminds me of the tale told by my high-school physics teacher, of a man jogging (or some such) in a lighning storm. A tree branch or something knocks a live powerline loose, which happens to hit him, sending his heart into fibrillations. As he was lying on the ground, almost assuredly going to die, lightning happened to strike an object nearby, whacking him with a nice, constant voltage. This "rebooted" his heart, essentially, just like a defibrillator at a hospital. It started beating normally again, and the guy was ok.
True story (according to my physics teacher).
Maybe I should RTFA, but it seems to me that the RFID tags would naturally be destroyed at the checkout counter -- that's what they do with current RF security tags (destroy or remove them). If it can't be done forcibly with an overload, then design them to respond to a "suicide" command.
If it's a concern that the chips will merely be disabled for purposes of letting them go out of the store, but still be active in other ways, I'd say a few lawsuits would cure any company of such desires. e.g. if Wal-Mart put them in shoes, all it takes is one advocate to buy a pair, cut out the RFID tag, and prove that it's still active. Why would any company risk the embarassment?
Yeah, me too.
But of course, then I remember the possibility that the use of floodlights that far down can also be damaging: any light-sensing organs on any creatures down there are going to be extremely sensitive, and will likely be fried by the amount of light needed to take a picture (with standard equipment).
So yes, a picture of one of these critters swimming/floating in their native habitat would be nicer than seeing them dead on an examination tray. But it's not so nice to know that the picture shows a creature that has just been permanently blinded by the very fact that you are looking at it.
- sigh -
Can you name your local representative to the state assembly?
I know I can't. I don't think many can. And that's where the deeper problem lies. It's easy to blame the Governor; it's a lot harder to study who the assemblypersons are and figure out which ones (if any) really deserve our votes.
Better not wear this in prison.
Might want to try a reverse-osmosis system. A cheap one will run not much more than $100. Have to replace the carbon filters every 6 months, at abour $40 each time, and the reverse-osmosis filter itself every couple of years; not sure how much that is. Not too hard to set up, especially if your kitchen sink has some pre-made accessory holes.
Probably less expensive than your current approach sounds.
[Ooog walks into cave]
Bogg: Hi, Ooog! I see you have your rock with you.
Ooog: Yeah, I know, it's kinda heavy, but, well, you know, ya gotta carry something...
Bogg: Oh, I know! Believe me!
[Bogg picks up his rock]
Bogg: I've had this sucker for 3 days now. It really fills the void left when I threw my last rock into the river...
Ooog: What'd you do that for, anyways?
Bogg: Oh, well, you know, gotta throw something...
Ooog: Oh yeah, I hear ya!
Boog: Hey, can you believe they're charging three squirrels for a good rock nowadays?
Ooog: Oh, man, that sucks! That's even more than a good clay pot -- those are 2 squirrels, maybe 2 and a half max!
Bogg: Yeah, man, why are there so many idiots who would pay so much just for a rock?!
Ooog: Well, ya know, gotta carry something...
Bogg: Oh, yeah, I hear ya!
I was an intern at a not-to-be-named company back in the early 90's. One of the employees had a party at his house on the lake. At one point, he was showing off his nifty little powerboat to me and 2 other interns. We interns were in our day clothes. There were no seat belts, only little handhold straps. And the water was a bit choppy. As you might expect, we 3 were pretty nervous as he raced the boat around, pulling sharp turns, and bouncing us around on the water. We felt like we were about to fly out of the boat!
Then, at one point, he stopped the boat, and said "ok, who wants to take the wheel now?" Having had experience with jet skiis, which are similar but smaller watercraft, I volunteered. After I started getting the hang of it, it started to be really fun, and I even started intentionally trying to do little jumps off the wakes of other boats and such. I no longer felt in danger of flying out of the boat, despite the fact that the only thing holding me in was my grip on the wheel.
After we got back to shore, one of the other interns gave his opinion: "yeah, it was pretty scary, with him bouncing us all over the place like that, but dude, you were way worse! You were *trying* to jump those waves!!"
Moral of the story: when you're in the driver's seat, you can trust your intentions, and your ability to predict the outcome of your actions. When someone else is running things, it's a lot harder not to be nervous.
The author of the original paragraph is obviously perceiving the world in this manner. His government, whose actions he is apparently willing treat as a proxy of his own, can be seen as "trusted'; others cannot. This is someone who needs to take the time to consider the other side of this symmetry.
Just an off-topic note, we didn't mess up in Somalia until we did two things:
They are.
Just have to get the actual material first. Only have aout 1.5 orders of magnitude to go before the nanotube-epoxy composite is strong enough...
Why isn't it NASA's main thrust yet? Gotta prove that it's possible first, with a real material.
> This worm appears to primarily affect Microsoft systems
<Nelson>
Ha - Haah!
</Nelson>
And now...
<Hanz&Franz>
Once again, ha haa! I lauugh at you silly foolz, with your flabby Windowz and your buuggy virus-baiiting Outlook email reader. I sit here with my puuumped-up Linux system, and my maanly Mutt text-only mail reader, and I open up my spam and virus emails and lauugh again because they cannot haarm me!
Ha Haaaah!
</Hanz&Franz>