Point of Linus was, taking a 6 core CPU, and replacing 2 cores with more cache and more transistors per core should make almost anything on Desktop run faster.
There's an element of truth to this, but on the other hand, cache space is already big enough that it hits the law of diminishing returns. Yes, the biggest performance hits in current computing are cache misses. But cache misses are already unexpected events, and cache misses are of biggest concern to the user when there are lots of them at once -- ie when iterating through a large bit of data. Text searches on large documents in a complex format (eg MS Word). Making a global change to a large file. These are the situations where performance matters, and these are the situations where you're going to get cache misses. Torvalds dismisses photo editing as a task for "professional photographers", but our amateur cameras are taking phenomenally detailed pictures, and even making fairly simple edits is a compute-intensive task. He may be right, but he may equally be wrong.
Yes, but it can't really do much non-graphics parallel processing at the same time as rendering a game. As I understand it, a lot of problems with AI in modern AAA titles is down to the fact that they need the parallelism, but in AAA-land, graphics are king, and the AI guys don't get enough cycles to do a decent job.
Verification is the process of checking that software works correctly. The more complex the system, the more complex the process of verification. Rather unfair of the GP to throw that in as a single word after you explicitly said that you're not a computer scientist.
In essence, it's already done that way. A System-on-a-chip (SoC) typically has a couple of general-purpose cores, along with sound and video processors. In a full-sized PC, the graphics processing is usually taken to another chip -- in fact another circuit board entirely. Because most of the work the graphics processor (=GPU) does is largely independent of the main processor (=CPU) (the CPU pushes in the data, says "do X with it", the GPU then churns away through the data) it doesn't need to be closely linked or share a lot of memory. In fact, it's more efficient for them not to share memory, as then they're not getting in each other's way.
Expanding that system for more types of semi-general-purpose cores would get rather complicated.
The issue is not whether parallelism is uniformly better for all tasks. The question is, is parallelism better for some tasks. And as Torvalds points out, those tasks do exist (Graphics being an obvious one).
I think, as one of the quoted comments in the article said, that current programming languages have a lot to answer for in this debate. If we look at mathematics, the sum (i.e. sigma) and product operators are inherently parallel. Sadly our FOR and WHILE loops are not, as procedural iteration often relies on side-effects. It's no accident that Scala uses the mathematical basis of the functional programming paradigm as a foundation for massively scalable, parallelisable programming (even though it's a mind-twisting wreck of a language in many ways).
Every non-trivial program will have to crunch through large iterative processes at some point, and even if these are a small percentage of execution time, the reality is that most interactive systems have a lot of idle time anyway, and the delay for the user is only when the program gets stuck in a long iteration. So it follows that even if parallelism reduces the overall performance, it is of no consequence as the perceived performance by the user is improved.
But I have to take issue of Linus tone in which he downplays "graphics" as being a rather unimportant subset of computing tasks. It's not "graphics". It's "GRAPHICS". That's not a small outlier of a task.
His point isn't that graphics is a small thing, it's that the GPU already handles that, and that we therefore don't need much parallelism in the CPU. But when I think about games, I think about AI, and the AI has to operate in 3D space, so the AI obviously benefits from the same parallelism as the graphics. Do we steal GPU cycles to run our AI? No, because the way of the market is typically that games sell first and foremost on their looks. So we need more parallel grunt. Plus, of course, as AI has to handle multiple independent agents, AI is an inherently parallel task (multithreaded in concept, regardless of whether a particular game implements it in threads or not)
But going back to Scala and FP... A lot of the problems of memory locking are nicely sidestepped if you implement your code in FP: FP guarantees immutability of values: you cannot write to an existing value, so you don't need to have notions of "atomic" operations, and hence no need to lock in most circumstances. Caching becomes less of a pain, as nothing is ever going to change, so your cache value cannot be incorrect.
Some of the comments in the article refered to theoretical extra bugs due to having to think in parallel, but it simply gives more motivation for a programming paradigm that is less bug-prone, and FP is that paradigm. FP has been rejected by programmers far too long, but the simple mechanism of immutability removes that most bothersome of bugs -- the erroneously altered value that you spend a week tracking down. FP should already be easier to reason about than procedural programming, if we learned to do it properly. FP for parallel isn't all that much different from single-threaded FP, so would actually make basic parallel code no more complicated to learn than algebra.
And how is that any better than playing to the rules? The official time for trading is before rolling the dice for your turn, and quite often traded prices do exceed face value. What's the benefit in your version?
If someone starts at 8 and another starts at 10, then from 10-12 there will be two desks occupied. Then one's at lunch from 12-1, and the other from 1-2 and from 2-4 both are at desks, and the second's in the office to 6. So even if that says there's only 70% occupancy, there is still no free desk for anybody. If you want to add coffee breaks, you're implying that every time you stand up, you're going to have to find another seat when you come back, then you're going to wait a while as your roaming profile downloads to the new PC. (True network workstations are still a minority thing.) And if you want 100% occupancy, that means people queuing for a seat -- more lost productivity. Seat-hours are less of a fungible commodity than even the mythical man-month.
When my former employer switched to hotdesking, anyone starting any later than 9 o'clock would spend half-an-hour to an hour looking for a seat. I got onto a support service and had a fixed desk, so I was alright....
What? That's a ridiculous comparison. If someone was to send you a parcel of cocaine and you got arrested because the police did a random check that day, then you'd agree that the guy who sent the drugs screwed you over. This is contraband. Sending it over risks getting the recipients killed.
And second, regardless of intention, science IS anti-religion. Science is based on logic and rationality, which must reject religion, since religion is based on faith (believing without proof or despite proof of the contrary). People who say that science and religion can go together either don't understand science, religion, or is just trying to make a square peg fit a round hole.
I would suggest that it is you that doesn't understand science or religion.
The university system was instituted by the Catholic church, with the aim of studying the mechanisms of the universe (hence the name). Muslim scholars of old were pivotal in the further development of Greek mathematical thought and of Greek and Middle Eastern astronomy. For most of history, studying the physical world has been considered a sacred endeavour -- under the religious viewpoint, this is studying the works of $DEITY.
As for logic and rationality, the basic concept of a supernatural deity sits outside of all repeatable observable evidence. You cannot make a logical conclusion either way without solid evidence. All we have is a few documents purporting to record witness testimony from centuries past. The most likely explanation for this witness evidence is a mixture of schizophrenic disorders and hallucinogens (in particular the fungus ergot), in my opinion, but I cannot state that Moses didn't actually see a burning bush. Even if we can discount certain events (eg the Great Flood), that does not mean that the existence of the related deity (in this case the god of Judaism, Christianity and Islam).
The only truly rational stance is agnosticism: I do not know, and cannot know, and because of this, that knowledge is irrelevant.
That is my stance, and I do not think it shows any intellectual integrity to mock somebody for believing in something that is not demonstrably false. Consider the recent "discovery" of the phenomenon called the "rogue wave" -- people had been talking about it for centuries, but this testimony was discounted as the same sort of fantasy that made sailors come up with stories about giant squid... which also turned out to be true. But of course not all witnesses were reliable, and we can be pretty sure there was never such a thing as a mermaid.
I personally believe that overly aggressive atheists have done more to harm science's standing within the religious community than the crazy fundamentalist preachers that nobody was listening to in many parts of the world many years ago.
You see, we disagree completely here. You seem to think that science keeping a good standing with religion is a good thing. I, on the other hand, think it is a very bad thing, because it can be used to lend validity to religion. Religion is mysticism, lies and control. It is (as proved by Ron Hubbard) a good way to get rich. It is a way to control people and to impose bronze age morals to a society much more evolved morally.
The brainwashing cult relies on its victims having limited reasoning powers. A broad education allows adherents of a religion to think for themselves. And if they can't think for themselves, how are they ever going to question their beliefs? Many religions want their followers to question (ie critically appraise) their beliefs. And I'm absolutely sure that you want religious people to question their beliefs. But you cannot arm them with the tools to question their beliefs if you tell them the tool is a poison that will kill their beliefs.
It is because preachers and churches in general started noticing they had to impose a firmer grip, because they were losing the battle against reason.
That would imply it was the same preachers and churches, but they are new preachers and churches. The change isn't the message the followers receive in the same old place, but the place they go to. Why did they go there? Did they go their seeking to be controlled? Do you genuinely b
But when you talk about "zombie-Jesus", you are not merely offending those who were offended by the tweet - you are offending the vast majority of Christians' and at risk of supporting the notion that "science" is anti-religion.
I personally believe that overly aggressive atheists have done more to harm science's standing within the religious community than the crazy fundamentalist preachers that nobody was listening to in many parts of the world many years ago. When I was younger, these extreme sects were restricted mostly to certain states of the US, and also parts of West Africa. They have spread alarmingly quickly. Why? Why did people turn away from the mainstream churches and turn fringe beliefs into something approaching a new orthodoxy? The preachers are getting a receptive flock from the get-go because secular media has already told them that science isn't for the religious.
You seem to be assuming I'm a religious person. But when you assume, you make an ass of U and... U alone. I did not once claim that the tweet was offensive. But clearly some folk think it is, and if you want to convince them otherwise, being deliberately offensive is not a particularly good strategy. If on the other hand, all you want to do is feel smug and superior...
Hmm... if you're trying to suggest people have no right to be offended, you should try to act like a reasonable individual who doesn't troll the religious. You will not win anyone over with talk of "zombie-jesus" - that's trolling, pure and simple. Trolling while sticking up for someone against accusations of trolling is where the old saying "with friends like these, who needs enemies" comes in....
Unfortunately, a great many scientists do end up believing their own myths later in their careers. A great many new theories were sat of for decades because they couldn't gain acceptance until the old guard died out, regardless of how much data they had to back it up. Ironically, one of the big examples used to show the antagonism of Christianity and science, the sentencing of Galileo Galilei, was just another example of one academic challenging the academic orthodoxy, but the "old guard" in that instance included the Pope and most of his cardinals -- the most powerful men on the planet.
What everyone forgets about dates is that the solstice is the 21st in the Gregorian calendar. Due to the inaccuracies of the Julian calendar, important solar events were unstable. The pegging to the 25th in the Gregorian calendar was a consequence of that. I don't know exactly why it happened, but comparing dates on the Gregorian calendar alone makes no sense.
Slashdot Mirror
Point of Linus was, taking a 6 core CPU, and replacing 2 cores with more cache and more transistors per core should make almost anything on Desktop run faster.
There's an element of truth to this, but on the other hand, cache space is already big enough that it hits the law of diminishing returns. Yes, the biggest performance hits in current computing are cache misses. But cache misses are already unexpected events, and cache misses are of biggest concern to the user when there are lots of them at once -- ie when iterating through a large bit of data. Text searches on large documents in a complex format (eg MS Word). Making a global change to a large file. These are the situations where performance matters, and these are the situations where you're going to get cache misses. Torvalds dismisses photo editing as a task for "professional photographers", but our amateur cameras are taking phenomenally detailed pictures, and even making fairly simple edits is a compute-intensive task. He may be right, but he may equally be wrong.
Yes, but it can't really do much non-graphics parallel processing at the same time as rendering a game. As I understand it, a lot of problems with AI in modern AAA titles is down to the fact that they need the parallelism, but in AAA-land, graphics are king, and the AI guys don't get enough cycles to do a decent job.
Verification is the process of checking that software works correctly. The more complex the system, the more complex the process of verification. Rather unfair of the GP to throw that in as a single word after you explicitly said that you're not a computer scientist.
In essence, it's already done that way. A System-on-a-chip (SoC) typically has a couple of general-purpose cores, along with sound and video processors. In a full-sized PC, the graphics processing is usually taken to another chip -- in fact another circuit board entirely. Because most of the work the graphics processor (=GPU) does is largely independent of the main processor (=CPU) (the CPU pushes in the data, says "do X with it", the GPU then churns away through the data) it doesn't need to be closely linked or share a lot of memory. In fact, it's more efficient for them not to share memory, as then they're not getting in each other's way.
Expanding that system for more types of semi-general-purpose cores would get rather complicated.
The issue is not whether parallelism is uniformly better for all tasks. The question is, is parallelism better for some tasks. And as Torvalds points out, those tasks do exist (Graphics being an obvious one).
I think, as one of the quoted comments in the article said, that current programming languages have a lot to answer for in this debate. If we look at mathematics, the sum (i.e. sigma) and product operators are inherently parallel. Sadly our FOR and WHILE loops are not, as procedural iteration often relies on side-effects. It's no accident that Scala uses the mathematical basis of the functional programming paradigm as a foundation for massively scalable, parallelisable programming (even though it's a mind-twisting wreck of a language in many ways).
Every non-trivial program will have to crunch through large iterative processes at some point, and even if these are a small percentage of execution time, the reality is that most interactive systems have a lot of idle time anyway, and the delay for the user is only when the program gets stuck in a long iteration. So it follows that even if parallelism reduces the overall performance, it is of no consequence as the perceived performance by the user is improved.
But I have to take issue of Linus tone in which he downplays "graphics" as being a rather unimportant subset of computing tasks. It's not "graphics". It's "GRAPHICS". That's not a small outlier of a task.
His point isn't that graphics is a small thing, it's that the GPU already handles that, and that we therefore don't need much parallelism in the CPU. But when I think about games, I think about AI, and the AI has to operate in 3D space, so the AI obviously benefits from the same parallelism as the graphics. Do we steal GPU cycles to run our AI? No, because the way of the market is typically that games sell first and foremost on their looks. So we need more parallel grunt. Plus, of course, as AI has to handle multiple independent agents, AI is an inherently parallel task (multithreaded in concept, regardless of whether a particular game implements it in threads or not)
But going back to Scala and FP... A lot of the problems of memory locking are nicely sidestepped if you implement your code in FP: FP guarantees immutability of values: you cannot write to an existing value, so you don't need to have notions of "atomic" operations, and hence no need to lock in most circumstances. Caching becomes less of a pain, as nothing is ever going to change, so your cache value cannot be incorrect.
Some of the comments in the article refered to theoretical extra bugs due to having to think in parallel, but it simply gives more motivation for a programming paradigm that is less bug-prone, and FP is that paradigm. FP has been rejected by programmers far too long, but the simple mechanism of immutability removes that most bothersome of bugs -- the erroneously altered value that you spend a week tracking down. FP should already be easier to reason about than procedural programming, if we learned to do it properly. FP for parallel isn't all that much different from single-threaded FP, so would actually make basic parallel code no more complicated to learn than algebra.
And how is that any better than playing to the rules? The official time for trading is before rolling the dice for your turn, and quite often traded prices do exceed face value. What's the benefit in your version?
If someone starts at 8 and another starts at 10, then from 10-12 there will be two desks occupied. Then one's at lunch from 12-1, and the other from 1-2 and from 2-4 both are at desks, and the second's in the office to 6. So even if that says there's only 70% occupancy, there is still no free desk for anybody. If you want to add coffee breaks, you're implying that every time you stand up, you're going to have to find another seat when you come back, then you're going to wait a while as your roaming profile downloads to the new PC. (True network workstations are still a minority thing.) And if you want 100% occupancy, that means people queuing for a seat -- more lost productivity. Seat-hours are less of a fungible commodity than even the mythical man-month.
When my former employer switched to hotdesking, anyone starting any later than 9 o'clock would spend half-an-hour to an hour looking for a seat. I got onto a support service and had a fixed desk, so I was alright....
What? That's a ridiculous comparison. If someone was to send you a parcel of cocaine and you got arrested because the police did a random check that day, then you'd agree that the guy who sent the drugs screwed you over. This is contraband. Sending it over risks getting the recipients killed.
And second, regardless of intention, science IS anti-religion. Science is based on logic and rationality, which must reject religion, since religion is based on faith (believing without proof or despite proof of the contrary). People who say that science and religion can go together either don't understand science, religion, or is just trying to make a square peg fit a round hole.
I would suggest that it is you that doesn't understand science or religion.
The university system was instituted by the Catholic church, with the aim of studying the mechanisms of the universe (hence the name). Muslim scholars of old were pivotal in the further development of Greek mathematical thought and of Greek and Middle Eastern astronomy. For most of history, studying the physical world has been considered a sacred endeavour -- under the religious viewpoint, this is studying the works of $DEITY.
As for logic and rationality, the basic concept of a supernatural deity sits outside of all repeatable observable evidence. You cannot make a logical conclusion either way without solid evidence. All we have is a few documents purporting to record witness testimony from centuries past. The most likely explanation for this witness evidence is a mixture of schizophrenic disorders and hallucinogens (in particular the fungus ergot), in my opinion, but I cannot state that Moses didn't actually see a burning bush. Even if we can discount certain events (eg the Great Flood), that does not mean that the existence of the related deity (in this case the god of Judaism, Christianity and Islam).
The only truly rational stance is agnosticism: I do not know, and cannot know, and because of this, that knowledge is irrelevant.
That is my stance, and I do not think it shows any intellectual integrity to mock somebody for believing in something that is not demonstrably false. Consider the recent "discovery" of the phenomenon called the "rogue wave" -- people had been talking about it for centuries, but this testimony was discounted as the same sort of fantasy that made sailors come up with stories about giant squid... which also turned out to be true. But of course not all witnesses were reliable, and we can be pretty sure there was never such a thing as a mermaid.
I personally believe that overly aggressive atheists have done more to harm science's standing within the religious community than the crazy fundamentalist preachers that nobody was listening to in many parts of the world many years ago.
You see, we disagree completely here. You seem to think that science keeping a good standing with religion is a good thing. I, on the other hand, think it is a very bad thing, because it can be used to lend validity to religion. Religion is mysticism, lies and control. It is (as proved by Ron Hubbard) a good way to get rich. It is a way to control people and to impose bronze age morals to a society much more evolved morally.
The brainwashing cult relies on its victims having limited reasoning powers. A broad education allows adherents of a religion to think for themselves. And if they can't think for themselves, how are they ever going to question their beliefs? Many religions want their followers to question (ie critically appraise) their beliefs. And I'm absolutely sure that you want religious people to question their beliefs. But you cannot arm them with the tools to question their beliefs if you tell them the tool is a poison that will kill their beliefs.
It is because preachers and churches in general started noticing they had to impose a firmer grip, because they were losing the battle against reason.
That would imply it was the same preachers and churches, but they are new preachers and churches. The change isn't the message the followers receive in the same old place, but the place they go to. Why did they go there? Did they go their seeking to be controlled? Do you genuinely b
If there really was a nation named Nerdonia we would bomb it to smithereens, invade it, exterminate all the inhabitants and incinerate anything left.
I don't think skin and hair grease are a substitute for crude oil.
But when you talk about "zombie-Jesus", you are not merely offending those who were offended by the tweet - you are offending the vast majority of Christians' and at risk of supporting the notion that "science" is anti-religion.
I personally believe that overly aggressive atheists have done more to harm science's standing within the religious community than the crazy fundamentalist preachers that nobody was listening to in many parts of the world many years ago. When I was younger, these extreme sects were restricted mostly to certain states of the US, and also parts of West Africa. They have spread alarmingly quickly. Why? Why did people turn away from the mainstream churches and turn fringe beliefs into something approaching a new orthodoxy? The preachers are getting a receptive flock from the get-go because secular media has already told them that science isn't for the religious.
You seem to be assuming I'm a religious person. But when you assume, you make an ass of U and... U alone. I did not once claim that the tweet was offensive. But clearly some folk think it is, and if you want to convince them otherwise, being deliberately offensive is not a particularly good strategy. If on the other hand, all you want to do is feel smug and superior...
Yes, I mean that the dates of solar events were unstable under the Julian calendar.
Hmm... if you're trying to suggest people have no right to be offended, you should try to act like a reasonable individual who doesn't troll the religious. You will not win anyone over with talk of "zombie-jesus" - that's trolling, pure and simple. Trolling while sticking up for someone against accusations of trolling is where the old saying "with friends like these, who needs enemies" comes in....
But if Pluto was bigger, it would actually be a planet.
Actually, no. It would need to have a less eccentric orbit.
That's racism against Romans.
Unfortunately, a great many scientists do end up believing their own myths later in their careers. A great many new theories were sat of for decades because they couldn't gain acceptance until the old guard died out, regardless of how much data they had to back it up. Ironically, one of the big examples used to show the antagonism of Christianity and science, the sentencing of Galileo Galilei, was just another example of one academic challenging the academic orthodoxy, but the "old guard" in that instance included the Pope and most of his cardinals -- the most powerful men on the planet.
What everyone forgets about dates is that the solstice is the 21st in the Gregorian calendar. Due to the inaccuracies of the Julian calendar, important solar events were unstable. The pegging to the 25th in the Gregorian calendar was a consequence of that. I don't know exactly why it happened, but comparing dates on the Gregorian calendar alone makes no sense.
Do you consider news stories to be "hamstering" of the preceding headline?
The masses who choose to read his stuff still count as masses. Nobody is forcing you to read his post.
Lemmiwinks? Is that you?
Yeah. No-one will ever need anything that COBOL can't do. All these modern languages are just ego.
Plug-ins are too platform specific. Implement it in JS for a "run anywhere" solution.
The best solution is always dependent on the task and the dataset.
(Obviously I mean updates, not insertions.)