Yeah and I call bullshit on that -- writing stuff down is just manual labour, pure and simple.
I made it through Vector calculus mostly without studying (I then tutored calc 1-3 for paying students). The act of paying attention and taking notes was sufficient to get it all into my head with understanding. I did occasionally read my notes or the book, but the writing is what worked for me. That's just me, everyone has their way of learning. I learn by doing, and taking notes is close enough to doing I guess. Another professor I had gave out printed course notes with each lecture. They had key pieces missing, so you had to pay attention to fill in the blanks, but the writing workload was significantly reduced so you could think about what it meant more. That also seemed to work well - it's like taking notes without the distraction.
The Bulldozer FPU is shared between the two cores in a module. It sounded like it's really only a single FPU for the new 256 bit instructions. Remember when the K8 used a 64bit FPU and still kicked Intels butt? IMHO they waited about a year too long to upgrade it to 128bit. Do we really know where BD bottleneck is?
It can't be easy to determine where the cuts are going to be made if you've decided to not just do an even across-the-board cut. Most divisions within any sizable company could have good arguments made for not selecting them for cuts.
AMD is a Fabless chip company now. That means they design chips. They are behind in performance on the x86 side, are about to be behind in low power when Intel uses FinFETs (sorry Tri-Gate). The last thing they need to cut is their core design business - it's what the company does.
It's because the temperatures recorded in urban areas are compensated for the extra warming.
Thanks for that. It was not clear to me that the data had been altered. In that case, one should simply discard all the urban data all the time since it's been forced to match the rural data. Makes one wonder what other manipulations are done to all this data?
Because it is assumed everyone these days possesses a GPU capable of doing the job.
That's probably a good assumption. However in the Linux world they're also assuming good GPU drivers which is not so much the case, but that is getting better. My NV44 is only recently supported well enough. My Intel EeePC is only recently getting there too. Next time I'll try AMD graphics and see how that goes.
The simplest pancake sorting algorithm requires at most 2n - 3 flips
That was my first thought. It took a bit to realize the problem is not sorting the pancakes, but finding the optimal sequence of flips to sort the pancakes.
AMD is now a fabless company. That means what it does is DESIGN CHIPS. When Hector took over, the K8 design was so far along all he did was focus on marketing it. That's an easy job when your new design is better than the competition. He didn't focus on maintaining fab capability and AMD is now a full generation behind Intel. He didn't focus on new design, so AMD is now behind Intel. So now that they're fabless, the only thing they do is design (and drivers for the GPU side). So this new guy apparently thinks they can just sell the existing fresh designs for a while. They mentioned cutting costs to go in a new direction - low power - but ANY direction they want to go will require design because that's what they do. On the other hand, if they want to just bolt ARM cores onto their GPUs the cuts make sense. I don't think going with ARM makes sense, that would be outsourcing half of their design business, or just the low power end of the CPU design business.
If they have next generation designs on track that will beat Intel in performance, and another to beat ARM in the low power areas, then they might feel free to *consider* cuts. Otherwise they are certainly killing themselves for short term gain.
I've been waiting years for the integration of:
1) CMU Sphinx
2) GNOME zeitgeist
3) Festival
4) A good looking virtual head with lip-sync to Festival
This would be the proper "office assistant". I want to log in and then ask the assistant (displayed instead of wallpaper) to bring up my email, or open the document I was working on yesterday, or whatever. Sure this would be rudimentary, but once the integration is done, people would actually start using these things and improving on them. Some would work on speech recognition - because they like it, but it's not good enough. Some would work on the AI - because it's neat, but needs to be smarter. Some would work on the assistant because they want photo-realistic hotness in their new virtual secretary.
First thing that comes to mind is: so what? This whole argument that smaller is better is crap. The reason that software is bigger these days is that it does more for you.
Smaller is usually better for a given functionality. Your point that software does more is valid too.
To interject with my own babble, why does a composited desktop require OpenGL? That stuff can be done with a few K of highly optimized C or ASM code. Another example of bloat in the name of easy to implement - never mind the added dependency.
So they thought heat islands - a documented phenomenon - would taint the data. They then go on to show that heat islands *don't* cause problems with the data - that the warming trend is the same weather you include data from the heat islands or not. So why is there no difference? This would seem to be a critical question to answer if we really want to understand climate change.
Just one more thing... It would be nice if you've got the CIO response in email. Then, should your whole world ever come crashing down due to lack of support (I don't think that's likely) and he blames YOU in front of his boss for skipping the support deal, you can just hold up the email. I'm being half sarcastic here - I don't think there will be a problem of that magnitude, and finger pointing at the CIO is only something you *might* do once you're screwed anyway and is really playing with fire. But having such things in a drawer is sometimes fun to think about;-)
So this guys model is not stable in the sense that it's own data can not be used to calibrate it. The natural conclusion is that one should run this test on their own models to determine how robust they are to their calibration procedure. If you can't produce the same model by calibrating it with it's own output, you've got a serious problem. This guy assumes that all models suffer from this, but I'm not convinced by his single data point.
If a CEO thinks "making money" is what the company is all about, he should sell all the assets and turn it into a hedge fund. Once you acknowledge that you're going to make money by making some product, then by all means do a good job at making that product. Or providing a service or whatever.
I think a better question is how well it compares to ARM.
How long do we put up with dark matter
on
Ask The Bad Astronomer
·
· Score: 3, Interesting
How long do we have to put up with the notion of "Dark Matter"? Whenever I research this, I come back to the "galactic rotation problem" as the most solid evidence. This discrepancy between prediction and observation is clearly rooted in the prediction being wrong. Keplers Laws do not apply to stars in galaxies. Hand waving and incorrect use of Gauss's Law have been going on for decades and we need it to stop. Why do people keep looking for "new physics" when they don't fully understand the physics we have?
Maybe if he had to actually work for a living at a minimum wage job, he'd stop asking those with little to no money to give up their chance to be raised up.
Loans are not free money. Student will have to pay them back. Making the loans more available means there is more money available for student to go to school. More money available means higher prices. So you pay for it either way, but doing it without loans means the price will be lower. Does anyone have stats on number of college graduates AND price of tuition over the last 25 years? Ideally I suppose these would be per capita number of students and inflation adjusted tuition.
MIke Melville was not looking for a job. He build an airplane from plans (a Rutan design) and flew it out there to show Burt. Burt offered him a job, and that was that. No degree, he just demonstrated that he could do exactly what they were doing. That led him to be the first private astronaut. Sure others at scaled went to MIT or whatever. I sent a resume to Scaled (along with a number of other people I know) and I got a phone interview and the others didn't. Why? Probably because my resume is full of verbs from DOing stuff. Parent post is spot-on.
Does Brazil have a minimum wage as high as ours? Do they have to comply with EPA regulations? Do any countries we claim to have "free trade" have the same burdens? I'm not saying those regulations are bad, just pointing out that it's not really free trade because of them. This is part of the reason the jobs are leaving the US.
Not all forms of mercury are identical, the half life for ethylmercury is something like a week. Of course, people like you thrive in ignorance so I doubt you ever bothered to look into any of this.
So you know every path that ethylmercury takes once inside the body? I don't doubt that some gets out. What I do doubt is that anyone knows what percent gets out and can guarantee that none of it accumulates somewhere (which would be bad). The GP question is entirely valid, why put it in when we know it will be bad IF any of it is absorbed? The answer may be that the alternatives have issues too, but when you have the option to completely avoid a hazard you should take that option unless the alternatives are worse.
The (erroneous, and much disproved) argument made by the anti-vaccine crowd is, explicitly, that the thimerosal causes autism.
No, that's some anti-vaccine people saying that. The actual hypothesis that was proposed was that the combination vaccine for MMR somehow allowed one component to lodge in the gut, or get through to the brain - I don't recall the story, but it was MMR not Mercury. The obvious solution was to test by giving the vaccines separately. I don't know how that ever turned out.
Anyway, many in the anti-vaccination crowd conflate all the supposedly bad things associated with them, leading some to think mercury causes autism.
Thimerosal has not been used in vaccines in Western nations in decades
It might not be used in children, but it is used. They were offering free flu shots where I work, and I asked them if it contained Thimerosal. They said yes, I passed, I got the flu (H1N1 season) and now it turns out I may be better off (see articles about H1N1 possibly triggering broad flu immunity). But my point is that it is still used here in the US.
I'm going to start by agreeing with you on there no evidence of a mercury-autism link. I just want to counter this point:
The compound in vaccines is not retained by the way elemental mercury is, it has a halflife of 18 days (it's actually removed from the brain even faster than that) and it does not interact biologically the same way elemental mercury does. You may as well avoid salt since Sodium and Chlorine are both poisonous
I've seen that argument when people complain about mercury in vaccines. Oh, it leaves the bloodstream with such and so time. My question is "where does it go". Leaving the blood stream is not the same as leaving the body. Mercury accumulates somewhere else (fatty tissue?). Is there conclusive proof that the mercury in vaccines does not get stored, changed, or in any way contribute to making you dumber? Which you ack that some mercury does. To me it seems a bad idea to use it - there are other preservatives.
They show a nice picture of a weather station at an airport. How many do they have in cornfields in the middle of Kansas? Airports tend to be in urban areas with large growth, and we already know cities create bubbles of higher temperature around them. Any station in such an area is simply documenting population growth and not saying much about global change.
Your hypothesis is that the world's best climate science researchers all spontaneously had strokes and started doing really bad research for no reason, that all pointed in the same direction?
I remember in the 1980s seeing a show (NOVA perhaps?) where the talked about how all the big climate models tended to fall into a "white earth" scenario if left to run for a while. Clearly the best models at the time were wrong. I'm sure they're better now than then, but I'm also sure some of the parameters in the models have been tweaked to fit existing data (which is normal, I'm not complaining) but it doesn't mean it's a totally valid model that we should make global policy decisions on.
I'm still hanging my hat on air travel - there is growing evidence of this. If we stopped putting con-trails over the whole planet, things would change significantly in short order. Would that change the long term outlook? I don't know, neither do you, and neither do the scientists. That experiment only lasted a week.
Slashdot Mirror
I made it through Vector calculus mostly without studying (I then tutored calc 1-3 for paying students). The act of paying attention and taking notes was sufficient to get it all into my head with understanding. I did occasionally read my notes or the book, but the writing is what worked for me. That's just me, everyone has their way of learning. I learn by doing, and taking notes is close enough to doing I guess. Another professor I had gave out printed course notes with each lecture. They had key pieces missing, so you had to pay attention to fill in the blanks, but the writing workload was significantly reduced so you could think about what it meant more. That also seemed to work well - it's like taking notes without the distraction.
The Bulldozer FPU is shared between the two cores in a module. It sounded like it's really only a single FPU for the new 256 bit instructions. Remember when the K8 used a 64bit FPU and still kicked Intels butt? IMHO they waited about a year too long to upgrade it to 128bit. Do we really know where BD bottleneck is?
AMD is a Fabless chip company now. That means they design chips. They are behind in performance on the x86 side, are about to be behind in low power when Intel uses FinFETs (sorry Tri-Gate). The last thing they need to cut is their core design business - it's what the company does.
Thanks for that. It was not clear to me that the data had been altered. In that case, one should simply discard all the urban data all the time since it's been forced to match the rural data. Makes one wonder what other manipulations are done to all this data?
That's probably a good assumption. However in the Linux world they're also assuming good GPU drivers which is not so much the case, but that is getting better. My NV44 is only recently supported well enough. My Intel EeePC is only recently getting there too. Next time I'll try AMD graphics and see how that goes.
That was my first thought. It took a bit to realize the problem is not sorting the pancakes, but finding the optimal sequence of flips to sort the pancakes.
AMD is now a fabless company. That means what it does is DESIGN CHIPS. When Hector took over, the K8 design was so far along all he did was focus on marketing it. That's an easy job when your new design is better than the competition. He didn't focus on maintaining fab capability and AMD is now a full generation behind Intel. He didn't focus on new design, so AMD is now behind Intel. So now that they're fabless, the only thing they do is design (and drivers for the GPU side). So this new guy apparently thinks they can just sell the existing fresh designs for a while. They mentioned cutting costs to go in a new direction - low power - but ANY direction they want to go will require design because that's what they do. On the other hand, if they want to just bolt ARM cores onto their GPUs the cuts make sense. I don't think going with ARM makes sense, that would be outsourcing half of their design business, or just the low power end of the CPU design business.
If they have next generation designs on track that will beat Intel in performance, and another to beat ARM in the low power areas, then they might feel free to *consider* cuts. Otherwise they are certainly killing themselves for short term gain.
I've been waiting years for the integration of:
1) CMU Sphinx
2) GNOME zeitgeist
3) Festival
4) A good looking virtual head with lip-sync to Festival
This would be the proper "office assistant". I want to log in and then ask the assistant (displayed instead of wallpaper) to bring up my email, or open the document I was working on yesterday, or whatever. Sure this would be rudimentary, but once the integration is done, people would actually start using these things and improving on them. Some would work on speech recognition - because they like it, but it's not good enough. Some would work on the AI - because it's neat, but needs to be smarter. Some would work on the assistant because they want photo-realistic hotness in their new virtual secretary.
Smaller is usually better for a given functionality. Your point that software does more is valid too.
To interject with my own babble, why does a composited desktop require OpenGL? That stuff can be done with a few K of highly optimized C or ASM code. Another example of bloat in the name of easy to implement - never mind the added dependency.
So they thought heat islands - a documented phenomenon - would taint the data. They then go on to show that heat islands *don't* cause problems with the data - that the warming trend is the same weather you include data from the heat islands or not. So why is there no difference? This would seem to be a critical question to answer if we really want to understand climate change.
Just one more thing... It would be nice if you've got the CIO response in email. Then, should your whole world ever come crashing down due to lack of support (I don't think that's likely) and he blames YOU in front of his boss for skipping the support deal, you can just hold up the email. I'm being half sarcastic here - I don't think there will be a problem of that magnitude, and finger pointing at the CIO is only something you *might* do once you're screwed anyway and is really playing with fire. But having such things in a drawer is sometimes fun to think about ;-)
So this guys model is not stable in the sense that it's own data can not be used to calibrate it. The natural conclusion is that one should run this test on their own models to determine how robust they are to their calibration procedure. If you can't produce the same model by calibrating it with it's own output, you've got a serious problem. This guy assumes that all models suffer from this, but I'm not convinced by his single data point.
If a CEO thinks "making money" is what the company is all about, he should sell all the assets and turn it into a hedge fund. Once you acknowledge that you're going to make money by making some product, then by all means do a good job at making that product. Or providing a service or whatever.
I think a better question is how well it compares to ARM.
How long do we have to put up with the notion of "Dark Matter"? Whenever I research this, I come back to the "galactic rotation problem" as the most solid evidence. This discrepancy between prediction and observation is clearly rooted in the prediction being wrong. Keplers Laws do not apply to stars in galaxies. Hand waving and incorrect use of Gauss's Law have been going on for decades and we need it to stop. Why do people keep looking for "new physics" when they don't fully understand the physics we have?
Loans are not free money. Student will have to pay them back. Making the loans more available means there is more money available for student to go to school. More money available means higher prices. So you pay for it either way, but doing it without loans means the price will be lower. Does anyone have stats on number of college graduates AND price of tuition over the last 25 years? Ideally I suppose these would be per capita number of students and inflation adjusted tuition.
MIke Melville was not looking for a job. He build an airplane from plans (a Rutan design) and flew it out there to show Burt. Burt offered him a job, and that was that. No degree, he just demonstrated that he could do exactly what they were doing. That led him to be the first private astronaut. Sure others at scaled went to MIT or whatever. I sent a resume to Scaled (along with a number of other people I know) and I got a phone interview and the others didn't. Why? Probably because my resume is full of verbs from DOing stuff. Parent post is spot-on.
Does Brazil have a minimum wage as high as ours? Do they have to comply with EPA regulations? Do any countries we claim to have "free trade" have the same burdens? I'm not saying those regulations are bad, just pointing out that it's not really free trade because of them. This is part of the reason the jobs are leaving the US.
Yep, eating fish (especially from some inland lakes) on a daily basis is discouraged for that very reason.
Here's a link.
So you know every path that ethylmercury takes once inside the body? I don't doubt that some gets out. What I do doubt is that anyone knows what percent gets out and can guarantee that none of it accumulates somewhere (which would be bad). The GP question is entirely valid, why put it in when we know it will be bad IF any of it is absorbed? The answer may be that the alternatives have issues too, but when you have the option to completely avoid a hazard you should take that option unless the alternatives are worse.
No, that's some anti-vaccine people saying that. The actual hypothesis that was proposed was that the combination vaccine for MMR somehow allowed one component to lodge in the gut, or get through to the brain - I don't recall the story, but it was MMR not Mercury. The obvious solution was to test by giving the vaccines separately. I don't know how that ever turned out.
Anyway, many in the anti-vaccination crowd conflate all the supposedly bad things associated with them, leading some to think mercury causes autism.
It might not be used in children, but it is used. They were offering free flu shots where I work, and I asked them if it contained Thimerosal. They said yes, I passed, I got the flu (H1N1 season) and now it turns out I may be better off (see articles about H1N1 possibly triggering broad flu immunity). But my point is that it is still used here in the US.
I've seen that argument when people complain about mercury in vaccines. Oh, it leaves the bloodstream with such and so time. My question is "where does it go". Leaving the blood stream is not the same as leaving the body. Mercury accumulates somewhere else (fatty tissue?). Is there conclusive proof that the mercury in vaccines does not get stored, changed, or in any way contribute to making you dumber? Which you ack that some mercury does. To me it seems a bad idea to use it - there are other preservatives.
They show a nice picture of a weather station at an airport. How many do they have in cornfields in the middle of Kansas? Airports tend to be in urban areas with large growth, and we already know cities create bubbles of higher temperature around them. Any station in such an area is simply documenting population growth and not saying much about global change.
I remember in the 1980s seeing a show (NOVA perhaps?) where the talked about how all the big climate models tended to fall into a "white earth" scenario if left to run for a while. Clearly the best models at the time were wrong. I'm sure they're better now than then, but I'm also sure some of the parameters in the models have been tweaked to fit existing data (which is normal, I'm not complaining) but it doesn't mean it's a totally valid model that we should make global policy decisions on.
I'm still hanging my hat on air travel - there is growing evidence of this. If we stopped putting con-trails over the whole planet, things would change significantly in short order. Would that change the long term outlook? I don't know, neither do you, and neither do the scientists. That experiment only lasted a week.