I just spend $3200 (including tax) on a maxed-out 17" MacBook Pro. Call me crazy or dedicated or both.
But in my opinion, unless you're developing for iOS or MacOS X, the Mac is a poor platform for LOTS of different kinds of things. I'm a chip designer, for instance, and there is ZERO software for the Mac in this area. I have to run Windows in a VM just so I can synthesize for FPGAs.
The Mac also has a dearth of good code editors. On Linux, I really liked nEdit. It has everything, and it is intuitive (or at least I felt that way when I was using it). For Windows, I've enjoyed using Edit Plus and Ultraedit. But for the Mac, the editors generally just suck. Every single one of them has some kind of amazingly bad UI design flaw. For instance, I think it was TextWrangler where searching with wrap-around and search&replace in selection are mutually exclusive -- if you want to use hotkeys to do these things, you have to open the Find dialog and change settings to go back and forth, and the devs told me this was intentional, like it was a good thing. For a while there, I considered blogging about it, it was that frustrating.:)
Actually, of all of them, Smultron is my favorite. It's back in development. It's still buggy as hell, but the author seems to be willing to listen when I report bugs. It's lightweight, and the UI is simple, consistent, and intuitive. So, it's getting there. Give it a version or two.
For a lot of kinds of development, I just use the Mac as a client for some other machine. If I'm doing web development, I actually just mount files on a Linux server using SMB (because netatalk has problems and OSX doesn't support NFS well) and serve the web pages from there.
What the Mac IS good at is content development. Multimedia, documents, etc. If I want to make a presentation or diagram, I use Pages, Numbers, and Keynote, along with MacTex and OmniGraffle. There are quite a number of other net clients (like NetNewsWire and Adium) and other sorts of apps that are just wonderful.
You can use Netbeans and Eclipse on the Mac, but Java just doesn't integrate all that cleanly with other Mac apps. You can get used to it, though. But generally, the Mac just isn't so great for software development.
I do scientific computing where we regularly use virtual address spaces larger than 4GB. Not all of that is in the working set, of course, but it's often necessary to have that much mapped. One recent example is my leakage power and delay models for near-threshold circuits. I implemented the Markovic forumlas and found them to be too slow. My simulations would take days. So, I figured out the granularities I needed for voltage, power, and temperature, and I implemented those models as giant look-up tables. The leakage power model occupies 4GB of address space all by itself. I just mmap the file into the process and go. Now the simulations take only hours.
A friend of mine was on a dating site and started chatting with someone who seemed interesting. At first, just on the site. Then via emails. Even then, he was concerned that she might not be a real person, because, you know, that happens. So he checked the email headers. They appeared to be legit.
She claimed to have an undergraduate degree in medicine. Since I have some knowledge in that area, I helped him converse with her on that topic. She became subtly evasive, in that she never had anything intelligent to say about it and avoided the topic in discussions she initiated. But we persisted, kept linking to articles, talking about recent developments, stuff in JAMA, etc. It was kinda funny, because it was obvious that either she had lied about her medical background.
Eventually, she sent him a picture. First thing he did, of course, was to do a reverse image search. That immediately lead him to a porn site for someone with an entirely different name. Likely: this is a scam. Unlikely but plausible: those really are her pics, and she'd done porn to pay for school or something. The thing is this wasn't just a pic of the same person. It was the exact same picture. He decided to play along anyhow. What harm could it do to not jump to (obvious) conclusions when you don't have absolute proof? No need to make accusations or anything.
It wasn't long before 'she' asked for money. Her boss had demanded sex, and when she refused left her stranded in a foreign country. She needed money for a plane ticket home. He knew this was coming and had his own sob story prepared. His car was in the shop, and his boss hadn't paid him the last two pay periods, so he wasn't going to be able to make his credit card and rent payments.
That was the last email exchanged.
This friend of mine is actually a really great guy, and I felt bad for him that someone had tried to scam him. Of course, he was also smart enough to not get emotionally involved. But what's interesting is that the checking he did wasn't all that hard. Maybe interpreting an email header is hard. But a reverse image search engine? Upload an image, see matches. Maybe the hard part is knowing that it exists. Either way, he was smart enough to have just the right level of paranoia (or discrimination, more like it). The problem with other people is that they never learn basic critical thinking. Are they just stupid? Or did their parents and schools never bother to teach it? My parents taught me to be wary of strangers, without being rude about it. Not everyone is out to get you, but you will frequently encounter people who have no qualms about cheating you. Size people up before trusting them.
This guy who was cheated out of $200k... how gullible was he? Who just sends money to a stranger? I don't care how convincing a person seems. I'm not sending them money. Even if I DO meet them, I'm not writing a check for $200k. A well-meaning person wouldn't ask for it in the first place.
In about 1997, I met a girl on the internet. She was a medical student in London. (Sound familiar?) But actually, she knew TONS about medicine, as well as a lot about our shared interest in psychology. This one seemed legit. I had some other friends in the UK, so I decided to visit. At Gatwick, I was greeted by a very attractive girl of Indian descent. She was exactly what she claimed to be. Today, she's a research cardiologist at a major London hospital. Weird, huh? Of course... she never asked me for money.
Huh. I always thought that other galaxies were moving away from us due to a _combination_ of expansion AND objects moving at sublight speeds through space.
> evolutionary theory keeps changing; what's "true" one day is "false" the next.
Citation needed.
Are you serious? This is the way science works! Every time a scientist makes a new discovery that contradicts current theory, those things contradicted are "falsified". However, the day before the discovery, those things were believed to be most likely true.
As for religion not changing, there's a very nice comment in the Religiosity gene discussion explaining how the interpretation of scripture changed with the needs of society etc.
> I want to show her the science
You have my support.
> its misleading to say that "evolution is true"
How is that misleading?
Because science isn't a truth-seeking engine. We hope that its results are true, but it's really a system for refining models. "True" is misleading. This implies finality and certainty we don't have. "True as far as we know right now" is more accurate.
To summarize some of the things I remember off the top of my head:
The design is basically area-equivalent to the Nehalem designs, but they've made certain structures more space efficient to make room to enlarge others. For instance, they've made the branch predictor use fewer bits for the same prediction accuracy. This and other improvements have allowed them to increase critical structures that affect things like the instruction window size. The instruction window pertains to the number of decoded but not executed instructions out-standing. A larger instruction window allows you to (a) find more instruction-level parallelism because you're more likely to find independent instructions that can be executed simultaneously, and (b) absorb the effect of some high latency operations, like L2 cache misses -- you can effectively hide much of the latency by continuing to look for and perform unrelated work during the stall. In Nehalem and before, they had a structure that unified the reservation station, register file, and reorder buffer. Logically, this makes sense, but it also makes that area very power hungry, and you can never turn it off. In Sandy Bridge, they've split those structures, so they can be clock-gated separately. Also, instead of accumulating dependency results in the reservation station, they're stored in a single centralized physical register file, and pointers are held in the RS. This saves a lot of space, since now instructions traveling around the processor just need to carry the pointer. (This does add some latency and writing required to fetch those results from the RF when they're finally needed.)
It's explicitly stated that Sandy Bridge is not a major revolution in processor design. Compared to Nehalem, you might think of it representing a large collection of efficiency improvements that work together to make a processor that is faster (clock for clock efficiency) and more power efficient.
Many of these improvements lead to the larger instruction window. IMHO, this is a critical improvement. A Sun engineer once described modern processing as being a race between last-level cache misses. You have an L2 miss, and you quickly run out of work to do, and the processor stalls until that out-standing read arrives. Meanwhile, you've accumulated a hundred cycles or so of pending work, which gets blasted through, and execution continues perhaps a little while until you have another L2 miss. Processors like Nehalem can execute four or more instructions per cycle (peak), but the effective AVERAGE instructions per clock is less than 1. These high-latency L2 misses are primarily responsible for that. Besides adding on-die memory controllers, which reduces the latency, Sandy Bridge lengthens the instruction window so as to absorb more of that latency, so that stall time is less.
Are you serious? Are you really sure about those "many years learing about evolutionary theory. It seems quite intuitive to me"? That you have the hang of its absolute basics, foundations?
This isn't for me. I did mention that I want to teach this to my daughter. Perhaps I would find it helpful to show HER the basics as part of explaining the theory.
I may feel that I understand it, but it's another matter entirely to teach it to someone else, especially a young person who is going to be, by nature, more literal in their thinking.
Honestly, now your first post starts to look like "concern troll" false flag deal.
With all due respect, your response makes you look like you didn't read my original post.
And I stand by my statement that you can't have a theory without a system of observations for the theory to explain. Evolutionary theory does not exist in a vacuum, yet too many people want to teach it that way.
Because of the controversy (artificial or not), a top-down approach to Evolution is probably not the best route. Biology of living animals is a better place to start, with DNA, reproduction, gene replication, replication errors, and observable effects of natural selection. From there, we can make the leap that evolution affecting different individuals differently would lead to genetic diversity, which would lead to speciation, which would explain why so many different creatures have so many features in common -- common ancestry. There's DNA evidence for that, as well as a fossil record. Work your way far enough back, and we have a solid basis for explaining the origin of all living organisms. The next step is to hypothesize about what those specific ancestors were. We have a theory based on current organisms that would suggest common ancestry, which can be used to explain the fossil record. The fossil record also lends credence to the common ancestry example. Since we can't go back and observe this directly, we need to converge multiple lines of evidence and reasoning.
So, what I would like to see is some data on the DNA and fossil evidence. So...
Another poster had suggested the Wikipedia article on Evolution. It's really very good, that and the "less technical version" it points to.
I'm dubious of Dawkins because he has an overt agenda against anything he perceives as non-scientific. It's this kind of bias that I want to avoid, because it makes it too easy to make a distorted interpretation of the facts. Many facts make evolution "obvious". Let's stick to those and not get mired in things that don't support the argument. (Mind you, those other facts are interesting in their own right, for other reasons.)
I don't care how people want to downplay it. Dawkins is a militant atheist who wants to destroy all religion. He's as bad as the fundamentalist Christians who want to destroy all science (and any GOOD religion as well). There are some bad religions out there, but parts of some of them make a good basis for teaching ethics. A "scientific theory of ethics" is much too dry for practical application.
I think you're completely missing my point. But we're getting into philosophical grounds that most people don't think about much.
The main idea here is that "fact" and "truth" are two different entities. Philosophically, you can have a truth not based on any facts, although that tends to be rather shaky in a practical sense. But think about moral ideas. To quantify ethics in terms of facts isn't impossible, but it's damn difficult. For most people, "killing is wrong" is self-evident and doesn't require any arguments regarding the evolutionary origin of altruism.
You can also have facts that don't lead to any "truth" because the facts are disconnected and don't have any underlying unifying cause.
Mathematics is one area that deals in truth. 2+2=4, and you can get there based on a very tiny number of axioms. Mathematical proofs depend on perfection in argument, although we can get into debates on whether or not our axioms are true...
As for science, consider the Standard Model of quantum electrodynamics. This is one of the most accurately predictive and well-tested theories in all of science. Yet we KNOW it's not true, because it doesn't explain everything. It's a brilliant model and comes damn close, but it's been contradicted. That doesn't mean we should chuck it. But this just shows you that nothing in science is ever going to be perfect. Indeed, that's a vital part of science. As soon as science discovers "truth", then you've reached a dead end in that field... nothing more to discover.
I've spent many years learing about evolutionary theory. It seems quite intuitive to me. But it isn't intuitive to many people because it's unlike anything they observe normally. Among scientists, evolution isn't controvercial, but among others, it is. Therefore, others need more convincing. But telling them to "just believe because you're an idiot if you don't" is just religion. To most people, evolution vs. something else is just a war between factions. There's no science in it. And while religion remains relatively stable, evolutionary theory keeps changing; what's "true" one day is "false" the next. The way that evolution is taught is partly responsible for this controversy.
I'm sure it exists, but I've never been able to find it; there's something that would really help: An up-to-date complete treatise of all the basic evidence that demonstrates the foundations of evolutionary theory. Observations of microevolution in the lab, sequences of fossils and how they were dated and how we're certain that they're from the same lineage, numerous clear examples, multiple convergent lines of evidence (fossils vs. dna), etc. In science class, they don't teach this. They teach the end results of the science as though it were FACT, but it's NOT. It is a fact that it's a good theory, but the theory itself cannot be deemed fact.
I have a little girl, and I don't want to just tell her "evolution is true, and those creationists are idiots." I want to show her the science. Besides, its misleading to say that "evolution is true". Evolutionary fact observed in the lab is true. Evolutionary theory is a MODEL that we STRIVE to MAKE true and is the best model we currently have. If it were TRUE, we'd be done. No more to discover. Rather, it is a gradually improving approximation.
Looks like I'm not the only one who noticed this, but due to various other UI bugs, I can't read people's full comments. Anyhow, using slashdot is making my browser (Safari) burn massive CPU cycles. Probably some timed event that fires off WAY too often.
Dude, they don't need phones to do this. When I flew back from Brazil, these three Brazilians a few rows ahead spent the whole fucking night yapping away. I had earplugs in as far as they would go, but these people were so loud they still kept me awake. I had the attendant ask them to quiet down. They flatly refused.
I've reported dozens of Firefox bugs over the years. Although I'm primarily a chip designer, I have studied usability (HCI, etc.), and I have a background in testing as well. I know about making intuitive systems, and I've been trained to be more objective about it, rather than just complaining about what I don't like. When I report Firefox bugs, they may get ignored because they're understaffed, but I've never had one tell me flat out that I was wrong. I HAVE had Chrome devs just tell me I'm wrong. Does working for Google automatically make you arrogant?
Firefox is one of the most important FOSS projects in the world, but there just aren't enough people working on it. (I'd love to help, but I'm a chip designer, so it's kinda out of my area.) For instance, there is a really long-standing bugs that affects Mac users, where Firefox prevents the machine from going to sleep if it's been idle. That on top of its general tendency to burn CPU cycles when completely idle (no animations or apps) makes Firefox a menace to notebook batteries. (Not as bad as Flash, but close.) Part of the problem is that there are basically no Mac developers working on Firefox. But this is an issue for other platforms as well. To most people, Firefox is more interesting to complain about than to work on.
Ok, I've done it myself. I've conflated power with energy. But what we're really talking about here with these "low power" devices is energy efficiency. It doesn't matter what the power consumption is if it takes forever to finish the job. You can make an ARM and an Atom use the same power consumption. The problem with the Atom is that it'll take a lot longer to finish the computation. Another facet is what a processor does with idle time. ARM's philosophy has always been to get the work done quickly and then power off, while Intel's still working on bringing their "halt" power down to a minimum.
The thing I've always wondered, something I've never seen mentioned, is how they deal with dust. Ok, so the walls keep out the large chunks, but what do they do to keep from drawing small particulate matter into their servers? I assume that they have filters on the intake vents, but they'd have to he more substantial than the ones used in facilities with traditional air conditioning, which would be a somewhat more closed environment, where the hot air circulates through the cooling system on its way back into the servers.
Ok, not entirely worthless. Linear algebra is used in loads and loads HPC workloads, but Linpack as a benchmark is NOT prepresentative of a typical real-world HPC workload. It focuses on peak flops, leaving behind things like inter-node bandwidth and latency, which are crucial for many important, real scientific supercomputing tasks.
Our CSE department chair recently quotes an article he read. To paraphrase, we're heading to the point where computation is going to be basically free, and what costs all the energy will be moving the data around. This is true for several reasons. One is the recent trend towards near-threshold computing. Ultra-low voltage (i.e. 400mV, when 900mV is nominal Vdd) can save 100x on power. It costs us 10x on speed, but now we can pack in 100x as many nodes into the same power and cooling budget, allowing for a 10x increase in aggregate throughput. But this works best for highly parallel and communication-heavy workloads. Fortunately, for many important areas (bombs, climate simulations, astronomy, real-time raytracing), this is the case. And moreover, people are getting better at parallelizing work.
I just flew to Atlanta for a conference. On my flight out, they selected me for screening. Of course, there was no indication as to why I was selected, although ostensibly, this is because the selection is random. Because I didn't want to be X-rayed, I opted to get the pat-down. I also did it because I wanted to know what it was like and as a subtle protest against the whole process (although I didn't tell them that -- I was curtious). As it turned out, it wasn't as bad as some people have described it. The examiner was gentle and efficient. He did "meet resistance" in the crotch area, insofar as I felt the side of his hand come into contact, but it didn't bother me. For ME, it was a curious experience. However, I can clearly see why many other people would find it to be invasive. I wasn't personally offended, but I still haven't changed my mind about it being excessive from a civil rights standpoint. The examiner was dispassionate in a way similar to a doctor, but the big difference is that when I get examined in personal places by a doctor, it's because I elected well in advance to have it happen.
This system sounds like it was half-baked by a bunch of turkeys who don't think about the inevitable mechanical failures. At some point, one of these containers will jam in their conduits. Collision with another container will leave an unfortunately inedible buffet strewn around in the tunnels, except for the rats that live there. Then people will go hungry and find themselves having to eat the engineers who designed this monstrosity.
I'm all for progress, but we've seen systems like this before, on New York's Roosevelt island, where they use a pneumatic garbage disposal system. It's worked reasonably well for a long time but requires significant constant maintenance, and it's seriously showing its age, with an accelerating rate of breakdowns that will only be solved with a complete overhaul that won't happen in the foreseeable future. The same thing will happen to this food subway. It may work okay for a few decades, and then it'll fall apart.
Our CS department chair made an interesting comment the other day (although I think he was quoting someone else). When it comes to energy consumption in todays processors, computation is almost free. All the energy goes into moving data around.
I read the article and watched the video, and it's certainly interesting. But what I don't understand is why there aren't more 6-stroke engines out there. In the 6-stroke engine, there's one intake/compression cycle, followed by the usual power cycle, followed by another power cycle that gets power by heating air from the surrounding engine heat. Unlike 4-stroke engines that seem to not get better than about 30% energy efficiency, the 6-strokes can get closer to 50%. They're mechanically a bit more complex, but they needn't be less reliable.
I'm tired of looking at benchmark figures. I want blog articles on how these JIT engines work (differently) and WHY one or anther is faster. Or at least why the FF engine is faster than it used to be.
Slashdot Mirror
I just spend $3200 (including tax) on a maxed-out 17" MacBook Pro. Call me crazy or dedicated or both.
But in my opinion, unless you're developing for iOS or MacOS X, the Mac is a poor platform for LOTS of different kinds of things. I'm a chip designer, for instance, and there is ZERO software for the Mac in this area. I have to run Windows in a VM just so I can synthesize for FPGAs.
The Mac also has a dearth of good code editors. On Linux, I really liked nEdit. It has everything, and it is intuitive (or at least I felt that way when I was using it). For Windows, I've enjoyed using Edit Plus and Ultraedit. But for the Mac, the editors generally just suck. Every single one of them has some kind of amazingly bad UI design flaw. For instance, I think it was TextWrangler where searching with wrap-around and search&replace in selection are mutually exclusive -- if you want to use hotkeys to do these things, you have to open the Find dialog and change settings to go back and forth, and the devs told me this was intentional, like it was a good thing. For a while there, I considered blogging about it, it was that frustrating. :)
Actually, of all of them, Smultron is my favorite. It's back in development. It's still buggy as hell, but the author seems to be willing to listen when I report bugs. It's lightweight, and the UI is simple, consistent, and intuitive. So, it's getting there. Give it a version or two.
For a lot of kinds of development, I just use the Mac as a client for some other machine. If I'm doing web development, I actually just mount files on a Linux server using SMB (because netatalk has problems and OSX doesn't support NFS well) and serve the web pages from there.
What the Mac IS good at is content development. Multimedia, documents, etc. If I want to make a presentation or diagram, I use Pages, Numbers, and Keynote, along with MacTex and OmniGraffle. There are quite a number of other net clients (like NetNewsWire and Adium) and other sorts of apps that are just wonderful.
You can use Netbeans and Eclipse on the Mac, but Java just doesn't integrate all that cleanly with other Mac apps. You can get used to it, though. But generally, the Mac just isn't so great for software development.
Is wired down, or is the URL invalid? Can someone post a link that works?
(Mainly, I'm curious to know what he looks like at 80.)
I do scientific computing where we regularly use virtual address spaces larger than 4GB. Not all of that is in the working set, of course, but it's often necessary to have that much mapped. One recent example is my leakage power and delay models for near-threshold circuits. I implemented the Markovic forumlas and found them to be too slow. My simulations would take days. So, I figured out the granularities I needed for voltage, power, and temperature, and I implemented those models as giant look-up tables. The leakage power model occupies 4GB of address space all by itself. I just mmap the file into the process and go. Now the simulations take only hours.
A friend of mine was on a dating site and started chatting with someone who seemed interesting. At first, just on the site. Then via emails. Even then, he was concerned that she might not be a real person, because, you know, that happens. So he checked the email headers. They appeared to be legit.
She claimed to have an undergraduate degree in medicine. Since I have some knowledge in that area, I helped him converse with her on that topic. She became subtly evasive, in that she never had anything intelligent to say about it and avoided the topic in discussions she initiated. But we persisted, kept linking to articles, talking about recent developments, stuff in JAMA, etc. It was kinda funny, because it was obvious that either she had lied about her medical background.
Eventually, she sent him a picture. First thing he did, of course, was to do a reverse image search. That immediately lead him to a porn site for someone with an entirely different name. Likely: this is a scam. Unlikely but plausible: those really are her pics, and she'd done porn to pay for school or something. The thing is this wasn't just a pic of the same person. It was the exact same picture. He decided to play along anyhow. What harm could it do to not jump to (obvious) conclusions when you don't have absolute proof? No need to make accusations or anything.
It wasn't long before 'she' asked for money. Her boss had demanded sex, and when she refused left her stranded in a foreign country. She needed money for a plane ticket home. He knew this was coming and had his own sob story prepared. His car was in the shop, and his boss hadn't paid him the last two pay periods, so he wasn't going to be able to make his credit card and rent payments.
That was the last email exchanged.
This friend of mine is actually a really great guy, and I felt bad for him that someone had tried to scam him. Of course, he was also smart enough to not get emotionally involved. But what's interesting is that the checking he did wasn't all that hard. Maybe interpreting an email header is hard. But a reverse image search engine? Upload an image, see matches. Maybe the hard part is knowing that it exists. Either way, he was smart enough to have just the right level of paranoia (or discrimination, more like it). The problem with other people is that they never learn basic critical thinking. Are they just stupid? Or did their parents and schools never bother to teach it? My parents taught me to be wary of strangers, without being rude about it. Not everyone is out to get you, but you will frequently encounter people who have no qualms about cheating you. Size people up before trusting them.
This guy who was cheated out of $200k... how gullible was he? Who just sends money to a stranger? I don't care how convincing a person seems. I'm not sending them money. Even if I DO meet them, I'm not writing a check for $200k. A well-meaning person wouldn't ask for it in the first place.
In about 1997, I met a girl on the internet. She was a medical student in London. (Sound familiar?) But actually, she knew TONS about medicine, as well as a lot about our shared interest in psychology. This one seemed legit. I had some other friends in the UK, so I decided to visit. At Gatwick, I was greeted by a very attractive girl of Indian descent. She was exactly what she claimed to be. Today, she's a research cardiologist at a major London hospital. Weird, huh? Of course... she never asked me for money.
Huh. I always thought that other galaxies were moving away from us due to a _combination_ of expansion AND objects moving at sublight speeds through space.
> evolutionary theory keeps changing; what's "true" one day is "false" the next.
Citation needed.
Are you serious? This is the way science works! Every time a scientist makes a new discovery that contradicts current theory, those things contradicted are "falsified". However, the day before the discovery, those things were believed to be most likely true.
As for religion not changing, there's a very nice comment in the Religiosity gene discussion explaining how the interpretation of scripture changed with the needs of society etc.
> I want to show her the science
You have my support.
> its misleading to say that "evolution is true"
How is that misleading?
Because science isn't a truth-seeking engine. We hope that its results are true, but it's really a system for refining models. "True" is misleading. This implies finality and certainty we don't have. "True as far as we know right now" is more accurate.
There are a number of really good articles on the advances in Sandy Bridge. For instance:
http://www.realworldtech.com/page.cfm?ArticleID=RWT091810191937
http://www.anandtech.com/show/3922/intels-sandy-bridge-architecture-exposed
To summarize some of the things I remember off the top of my head:
The design is basically area-equivalent to the Nehalem designs, but they've made certain structures more space efficient to make room to enlarge others. For instance, they've made the branch predictor use fewer bits for the same prediction accuracy. This and other improvements have allowed them to increase critical structures that affect things like the instruction window size. The instruction window pertains to the number of decoded but not executed instructions out-standing. A larger instruction window allows you to (a) find more instruction-level parallelism because you're more likely to find independent instructions that can be executed simultaneously, and (b) absorb the effect of some high latency operations, like L2 cache misses -- you can effectively hide much of the latency by continuing to look for and perform unrelated work during the stall. In Nehalem and before, they had a structure that unified the reservation station, register file, and reorder buffer. Logically, this makes sense, but it also makes that area very power hungry, and you can never turn it off. In Sandy Bridge, they've split those structures, so they can be clock-gated separately. Also, instead of accumulating dependency results in the reservation station, they're stored in a single centralized physical register file, and pointers are held in the RS. This saves a lot of space, since now instructions traveling around the processor just need to carry the pointer. (This does add some latency and writing required to fetch those results from the RF when they're finally needed.)
It's explicitly stated that Sandy Bridge is not a major revolution in processor design. Compared to Nehalem, you might think of it representing a large collection of efficiency improvements that work together to make a processor that is faster (clock for clock efficiency) and more power efficient.
Many of these improvements lead to the larger instruction window. IMHO, this is a critical improvement. A Sun engineer once described modern processing as being a race between last-level cache misses. You have an L2 miss, and you quickly run out of work to do, and the processor stalls until that out-standing read arrives. Meanwhile, you've accumulated a hundred cycles or so of pending work, which gets blasted through, and execution continues perhaps a little while until you have another L2 miss. Processors like Nehalem can execute four or more instructions per cycle (peak), but the effective AVERAGE instructions per clock is less than 1. These high-latency L2 misses are primarily responsible for that. Besides adding on-die memory controllers, which reduces the latency, Sandy Bridge lengthens the instruction window so as to absorb more of that latency, so that stall time is less.
Are you serious? Are you really sure about those "many years learing about evolutionary theory. It seems quite intuitive to me"? That you have the hang of its absolute basics, foundations?
This isn't for me. I did mention that I want to teach this to my daughter. Perhaps I would find it helpful to show HER the basics as part of explaining the theory.
I may feel that I understand it, but it's another matter entirely to teach it to someone else, especially a young person who is going to be, by nature, more literal in their thinking.
Honestly, now your first post starts to look like "concern troll" false flag deal.
With all due respect, your response makes you look like you didn't read my original post.
And I stand by my statement that you can't have a theory without a system of observations for the theory to explain. Evolutionary theory does not exist in a vacuum, yet too many people want to teach it that way.
Because of the controversy (artificial or not), a top-down approach to Evolution is probably not the best route. Biology of living animals is a better place to start, with DNA, reproduction, gene replication, replication errors, and observable effects of natural selection. From there, we can make the leap that evolution affecting different individuals differently would lead to genetic diversity, which would lead to speciation, which would explain why so many different creatures have so many features in common -- common ancestry. There's DNA evidence for that, as well as a fossil record. Work your way far enough back, and we have a solid basis for explaining the origin of all living organisms. The next step is to hypothesize about what those specific ancestors were. We have a theory based on current organisms that would suggest common ancestry, which can be used to explain the fossil record. The fossil record also lends credence to the common ancestry example. Since we can't go back and observe this directly, we need to converge multiple lines of evidence and reasoning.
So, what I would like to see is some data on the DNA and fossil evidence. So...
Another poster had suggested the Wikipedia article on Evolution. It's really very good, that and the "less technical version" it points to.
I'm dubious of Dawkins because he has an overt agenda against anything he perceives as non-scientific. It's this kind of bias that I want to avoid, because it makes it too easy to make a distorted interpretation of the facts. Many facts make evolution "obvious". Let's stick to those and not get mired in things that don't support the argument. (Mind you, those other facts are interesting in their own right, for other reasons.)
I don't care how people want to downplay it. Dawkins is a militant atheist who wants to destroy all religion. He's as bad as the fundamentalist Christians who want to destroy all science (and any GOOD religion as well). There are some bad religions out there, but parts of some of them make a good basis for teaching ethics. A "scientific theory of ethics" is much too dry for practical application.
I think you're completely missing my point. But we're getting into philosophical grounds that most people don't think about much.
The main idea here is that "fact" and "truth" are two different entities. Philosophically, you can have a truth not based on any facts, although that tends to be rather shaky in a practical sense. But think about moral ideas. To quantify ethics in terms of facts isn't impossible, but it's damn difficult. For most people, "killing is wrong" is self-evident and doesn't require any arguments regarding the evolutionary origin of altruism.
You can also have facts that don't lead to any "truth" because the facts are disconnected and don't have any underlying unifying cause.
Mathematics is one area that deals in truth. 2+2=4, and you can get there based on a very tiny number of axioms. Mathematical proofs depend on perfection in argument, although we can get into debates on whether or not our axioms are true...
As for science, consider the Standard Model of quantum electrodynamics. This is one of the most accurately predictive and well-tested theories in all of science. Yet we KNOW it's not true, because it doesn't explain everything. It's a brilliant model and comes damn close, but it's been contradicted. That doesn't mean we should chuck it. But this just shows you that nothing in science is ever going to be perfect. Indeed, that's a vital part of science. As soon as science discovers "truth", then you've reached a dead end in that field... nothing more to discover.
Uh.... the "general" model of evolution has to be based on a inductive model of the specifics. Without the specifics, the generality doesn't exist!
Every heard of a battery?
I've spent many years learing about evolutionary theory. It seems quite intuitive to me. But it isn't intuitive to many people because it's unlike anything they observe normally. Among scientists, evolution isn't controvercial, but among others, it is. Therefore, others need more convincing. But telling them to "just believe because you're an idiot if you don't" is just religion. To most people, evolution vs. something else is just a war between factions. There's no science in it. And while religion remains relatively stable, evolutionary theory keeps changing; what's "true" one day is "false" the next. The way that evolution is taught is partly responsible for this controversy.
I'm sure it exists, but I've never been able to find it; there's something that would really help: An up-to-date complete treatise of all the basic evidence that demonstrates the foundations of evolutionary theory. Observations of microevolution in the lab, sequences of fossils and how they were dated and how we're certain that they're from the same lineage, numerous clear examples, multiple convergent lines of evidence (fossils vs. dna), etc. In science class, they don't teach this. They teach the end results of the science as though it were FACT, but it's NOT. It is a fact that it's a good theory, but the theory itself cannot be deemed fact.
I have a little girl, and I don't want to just tell her "evolution is true, and those creationists are idiots." I want to show her the science. Besides, its misleading to say that "evolution is true". Evolutionary fact observed in the lab is true. Evolutionary theory is a MODEL that we STRIVE to MAKE true and is the best model we currently have. If it were TRUE, we'd be done. No more to discover. Rather, it is a gradually improving approximation.
Looks like I'm not the only one who noticed this, but due to various other UI bugs, I can't read people's full comments. Anyhow, using slashdot is making my browser (Safari) burn massive CPU cycles. Probably some timed event that fires off WAY too often.
Dude, they don't need phones to do this. When I flew back from Brazil, these three Brazilians a few rows ahead spent the whole fucking night yapping away. I had earplugs in as far as they would go, but these people were so loud they still kept me awake. I had the attendant ask them to quiet down. They flatly refused.
People don't need phones to be rude.
I've reported dozens of Firefox bugs over the years. Although I'm primarily a chip designer, I have studied usability (HCI, etc.), and I have a background in testing as well. I know about making intuitive systems, and I've been trained to be more objective about it, rather than just complaining about what I don't like. When I report Firefox bugs, they may get ignored because they're understaffed, but I've never had one tell me flat out that I was wrong. I HAVE had Chrome devs just tell me I'm wrong. Does working for Google automatically make you arrogant?
Firefox is one of the most important FOSS projects in the world, but there just aren't enough people working on it. (I'd love to help, but I'm a chip designer, so it's kinda out of my area.) For instance, there is a really long-standing bugs that affects Mac users, where Firefox prevents the machine from going to sleep if it's been idle. That on top of its general tendency to burn CPU cycles when completely idle (no animations or apps) makes Firefox a menace to notebook batteries. (Not as bad as Flash, but close.) Part of the problem is that there are basically no Mac developers working on Firefox. But this is an issue for other platforms as well. To most people, Firefox is more interesting to complain about than to work on.
Ok, I've done it myself. I've conflated power with energy. But what we're really talking about here with these "low power" devices is energy efficiency. It doesn't matter what the power consumption is if it takes forever to finish the job. You can make an ARM and an Atom use the same power consumption. The problem with the Atom is that it'll take a lot longer to finish the computation. Another facet is what a processor does with idle time. ARM's philosophy has always been to get the work done quickly and then power off, while Intel's still working on bringing their "halt" power down to a minimum.
The thing I've always wondered, something I've never seen mentioned, is how they deal with dust. Ok, so the walls keep out the large chunks, but what do they do to keep from drawing small particulate matter into their servers? I assume that they have filters on the intake vents, but they'd have to he more substantial than the ones used in facilities with traditional air conditioning, which would be a somewhat more closed environment, where the hot air circulates through the cooling system on its way back into the servers.
Ok, not entirely worthless. Linear algebra is used in loads and loads HPC workloads, but Linpack as a benchmark is NOT prepresentative of a typical real-world HPC workload. It focuses on peak flops, leaving behind things like inter-node bandwidth and latency, which are crucial for many important, real scientific supercomputing tasks.
Our CSE department chair recently quotes an article he read. To paraphrase, we're heading to the point where computation is going to be basically free, and what costs all the energy will be moving the data around. This is true for several reasons. One is the recent trend towards near-threshold computing. Ultra-low voltage (i.e. 400mV, when 900mV is nominal Vdd) can save 100x on power. It costs us 10x on speed, but now we can pack in 100x as many nodes into the same power and cooling budget, allowing for a 10x increase in aggregate throughput. But this works best for highly parallel and communication-heavy workloads. Fortunately, for many important areas (bombs, climate simulations, astronomy, real-time raytracing), this is the case. And moreover, people are getting better at parallelizing work.
I just flew to Atlanta for a conference. On my flight out, they selected me for screening. Of course, there was no indication as to why I was selected, although ostensibly, this is because the selection is random. Because I didn't want to be X-rayed, I opted to get the pat-down. I also did it because I wanted to know what it was like and as a subtle protest against the whole process (although I didn't tell them that -- I was curtious). As it turned out, it wasn't as bad as some people have described it. The examiner was gentle and efficient. He did "meet resistance" in the crotch area, insofar as I felt the side of his hand come into contact, but it didn't bother me. For ME, it was a curious experience. However, I can clearly see why many other people would find it to be invasive. I wasn't personally offended, but I still haven't changed my mind about it being excessive from a civil rights standpoint. The examiner was dispassionate in a way similar to a doctor, but the big difference is that when I get examined in personal places by a doctor, it's because I elected well in advance to have it happen.
This system sounds like it was half-baked by a bunch of turkeys who don't think about the inevitable mechanical failures. At some point, one of these containers will jam in their conduits. Collision with another container will leave an unfortunately inedible buffet strewn around in the tunnels, except for the rats that live there. Then people will go hungry and find themselves having to eat the engineers who designed this monstrosity.
I'm all for progress, but we've seen systems like this before, on New York's Roosevelt island, where they use a pneumatic garbage disposal system. It's worked reasonably well for a long time but requires significant constant maintenance, and it's seriously showing its age, with an accelerating rate of breakdowns that will only be solved with a complete overhaul that won't happen in the foreseeable future. The same thing will happen to this food subway. It may work okay for a few decades, and then it'll fall apart.
Our CS department chair made an interesting comment the other day (although I think he was quoting someone else). When it comes to energy consumption in todays processors, computation is almost free. All the energy goes into moving data around.
Would those be the Time Lords or the Auditors of Reality?
I read the article and watched the video, and it's certainly interesting. But what I don't understand is why there aren't more 6-stroke engines out there. In the 6-stroke engine, there's one intake/compression cycle, followed by the usual power cycle, followed by another power cycle that gets power by heating air from the surrounding engine heat. Unlike 4-stroke engines that seem to not get better than about 30% energy efficiency, the 6-strokes can get closer to 50%. They're mechanically a bit more complex, but they needn't be less reliable.
So what's holding us back?
I'm tired of looking at benchmark figures. I want blog articles on how these JIT engines work (differently) and WHY one or anther is faster. Or at least why the FF engine is faster than it used to be.