The underlying foundation of XGL is exactly what a compositing window system needs --- a generic OpenGL stack that unifies control of the GPU into a single multclient-aware driver. Yes, it complicates the driver. The driver needs to properly handle and schedule multiple rendering clients, it must manage do good memory management and video memory virtualization, and it must properly handle synchronization of rendering. However, handling these things properly is The Right Solution. Aiglx just perpetuates the driver sharing lunacy that exists now, which is something that OS X and Vista will get rid off.
The reason the Mac UI works so well is that its eyecandy is a method of subtly including information that might otherwise be lost.
Your point is well-taken, but I'd suggest you sit down with a copy of OS X 10.0 sometime. The eye-candy was pretty unsubtle back then. The refinement present in Tiger took Apple several years to get right. XGL is not yet a year old. Give it some time to mature.
XGL isn't a rewrite of the server. It's a rewrite of the DDX (device-dependent) portion. That's probably the best part of the server to rewrite though, given the DIX (device-independent) is relatively clean code. XGL doesn't get rid of X's cruftiest part, though, which is xlib. XCB is ready to be a replacement, but GNOME won't be able to move to it until 3.x, because Xlib is implicitly a part of the GDK ABI.
That said, I wouldn't say XGL is better than OS X yet. OS X can do the effects you listed, it just doesn't do a lot of them for asthetic reasons. Technically, I'd argue OS X's approach is superior to XGL's, since Quartz 2D Extreme uses a direct-rendering model as opposed to XGL's indirect model. Additionally, the fact that the compositor is seperate from the window server in XGL makes synchronization a much bigger PITA than in OS X. On the other hand, the indirect model allows the X server to access the geometry stream, which allows some effects the direct-rendering model doesn't. Technical merits aside, OS X still wins because its already a stable, mature, and widely used technology. It'll be awhile before XGL is as mature as Quartz (especially at the driver layer --- DRI is really not ready for XGL yet), and before GNOME/KDE apps use vector graphics as widely as OS X apps do.
However, I agree that war sux and any civilian death is a travesty. Everyone in America, including the leadership and military, agrees. Can you say the same about the Arab world?
Frankly, nobody in the United States gives a shit about civilian death in the Middle East. Honestly, nobody here thinks twice about it. We've got our own problems, and we really don't care about theirs. Believe it or not, the Arab world is the same way. They really couldn't care less, because it doesn't affect them directly. You average Arab thinks about the death-toll resulting from terrorism about as often as your average grocery store worker thinks about the death-toll resulting from war in the Middle East.
Regarding 3: You should be aware that radical Muslim clerics make up whatever Islamic laws they want. Their view of Islamic law has no bearing on reality. You're absolutely right, though, that a tolerable Muslim is one who does not follow the practices of his faith to the letter. Of course, the same is true of Christians. A real honest-to-God follow every letter of the Bible Christian would be quite a monstrous person based on our modern standards.
But seriously, you're joking right? If Christians were like Muslims, all those prominent gay marriage ceremonies a while ago would have been fire-bombed. Or do you think if someone tried to have a prominent, public gay marriage in, say, ANY Muslim country, they would be allowed?
It's not a Christian versus Muslim thing here. It's a rich country versus poor country thing. I'd love to see how a prominent, public gay marriage would go over in 18th century peasant Russia. No firebombs, probably, with that level of technology, but pitchforks would abound.
The real problem is not that we're dealing with Islam, but rather that we're dealing with poor and poorly educated people from culturally-backwards parts of the world. Attacking Islam won't solve anything, it's just a diversion for those who want to take action but don't want to really think about what action will be effective.
There is something serious sick in Islam today and unless some of their leaders stand up and turn the ship around,
That's really part of the problem. Islam doesn't really have any leaders. There is nobody in charge, and no formal hierarchy. It's mostly just a bunch of mobs controlled by Mullahs with a limited range of influence.
I was talking to a friend the other day, and trying to figure out a good analogy for the rarity of acts of women raping men. So I say "a women rapist is about as common as... ", and my friend fills in "a black person who doesn't commit crime?" I thought it was pretty funny, and I know he meant it entirely in jest, but I know if I ever said that out-loud in down-town DC, I'd get my ass seriously kicked. Some cartoonist, living in a place with a large and recently-emigrated (read: poor and relativey uneducated) Muslim population, depicted Mohammed as a terrorist. What did he expect? I believe its his right to say what he wishes, but some actions are just ill-advised. If I (I'm India, btw) went into the heart of the Bible Belt and started spouting off what I really thought about Christian conservatives, do you really think I wouldn't get a few death threats?
I agree with you that Muslims need to grow a thicker skin, but pinning the blame on Islam is like saying that black people are easily offended because they are black. It completely takes the cultural element out of the equation. There are a lot of people who can't take a joke, from women to hispanics to blacks to Christians to Arabs. I'd prefer to believe the reasons are cultural rather than the result of their innate tendencies...
If a religion isn't useful at getting people to love one another and not murder one another, its a failure as a religion. While you can't necessarily judge a religion based upon the actions of its followers, you would have to be an idiot to not draw some empirical evidence from these violent events about the nature of the Islam religion.
Throughout history, religion has been used as a political tool. During the wars of religion in Europe, tens of thousands of Christians slaughtered each other on the basis of Protestantism versus Catholicism. 3 million people died in France's wars of religion. Was that not evil?
The basic thing you fail to realize is that civil unrest is caused by economic, social, and political tension. The Muslim nations have more than their share of these things. The reason Christian societies have outgrown such primitive forms of violence* is because they are, on the whole, mature and economically-sound societies. At times when Muslim societies have had economic parity with Christian ones, their level of civility was quite comparable.
The evidence of history strongly suggests that religion is mostly uncorrelated with the behavior of populations. The Eastern religions are a lot mellower than any of the Semitic ones, yet that never stopped the Chinese and Japanese from having very brutal empires. Heck, Christianity did nothing to mitigate the brutality of the colonial powers either, for that matter.
*) They embrace, of course, more sophisticated and subtle, but ultimately more destructive forms of violence like launching cruise missiles from thousands of miles away...
We know for a fact that Muhammed led troops in war and had multiple wives. Seems wrong to me.
The personal character of Mohammed is something of a red-herring. Mohammed is not divine, he is simply the first leader of the faith. His place in Islam is more comparable to the place of someone like Charlemagne in Christianity rather than Jesus. As for polygamy, that's an anachronistic argument. Sects of Christianity have accepted polygamy up to the last century, and at the time when Mohammed lived, polygamy was well-accepted. I should point out that the Bible doesn't actually make any condemnations of polygamy --- that is a practice that has become part of Christianity as it has evolved.
In Christianity, the man Jesus Christ (Iesous Christos in greek) spoke at great length about how his true disciples should be harmless as doves, turn the other cheek, give, love, and forgive... They are not following the example set by Christ and therefore are committing acts of evil.
You can't just pick and choose the parts of the Bible you like. Yes, Christ said a lot of good things. But Christianity isn't merely the teachings of Christ. It includes everything in the Bible, along with the various accumualted doctrines in the case of Catholicism. There is a lot of distasteful stuff in the rest of the Bible. At the same time, there are many good things in the Quran too, along with several distasteful things.
Killing? obviously evil. no one should do it.
So our soldiers in Iraq are comitting evil acts?
Does it mean I'm gonna go killing fags?
Damn it. Here I thought I was replying to a legitimate post, now I realize I'm just talking to a half-wit.
Two things that have always bothered me about islamic martyrs: 1) unlike Jesus, and Stephen, who were peaceful martyrs killed for their words, islamic martyrs are just suicide bombers.
In theory, Islamic martyrs are
2) the koran says that muslims who martyr themselves will get however many virgins right?
Many respected scholars view that as a mistranslation, pointing out that the original text makes more sense as saying they will recieve "sweet fruits and drink" than virgins. But you will of course believe whatever you want to believe.
but in islam, women are treated as subhuman and worthless, not allowed to have rights.
AMD hasn't changed sockets just for kicks. The 754 to 939 transition was to add extra pins for the dual-channel memory controller. The AM2 socket transition will be to add support for DDR2 memory. These things required not just extra pins, but extra traces on the motherboard. Moreover, the traces have different timing characteristics because of the change in memory type. So even if AMD had used a socket with extra pins, old motherboards still wouldn't have the right lines to connect them to.
What second taskbar? There is only one taskbar. The thing at the top is a menu bar. It's there because the top and bottom of the screen are easiest to access, so it makes sense to use them fully for important functions like launching and switching apps.
Now, for the confusing part. Why was their previous allocator so lame compared to malloc()?
Because glibc's malloc() is actually a pretty fast and scalable piece of code for a general-purpose memory allocator. Even GNOME's new special-purpose allocator only gets about twice the performance of glibc's.
There is really not a whole lot of benefit in making the rockets reusable. The current SRBs are solid fuel boosters, which means they are basically just metal shells filled with a propellent/oxidizer mix. The metal shells are reused, but even if they weren't the savings wouldn't be significant.
To use your analogy, it'd be like a book coming with a sticker that states "You must be able to read English to use this book!" Yes, you can buy any book you like, but you can't complain too loudly when you buy a book you can't read. The publisher doesn't have to make it available in your language.
Nobody is asking Apple to make OS X available on generic PCs. They are just saying they have no power to sue if people manage to make the OS run on generic PCs. Your analogy is thus incorrect. It's more like the book comes with a sticker saying "you must read English to read this book", and the publisher sues you if have a friend translate it into Russian so you can read it.
The GPL is not a useage contract, it is a copyright license. It governs redistribution of a work, something that is entirely within its power as a copyright license. Apple's EULA is supposedly a contract, but the legal arguments for "opt out" contracts are not exactly solid. Moreover, contract law is a large and complex subject, and it is not at all clear that Apple is allowed to set the kind of conditions they do in their EULA. If a publisher had a shrink-wrapped EULA saying that you could only keep that book on specially-approved shelves, it is very probable that suits over violations of that EULA would be thrown out of court.
That said, at the moment it is not possible to use OS X on Intel without violating Apple's copyright. Since the OS X 10.4.4 binaries are only available via purchase of a new Mac, and the Mac itself comes with a copy of the OS installed, the only way to use OS X without violating Apple's copyright by copying the binaries is to delete it from the Mac and use the restore disc on another machine. When shrink-wrapped copies of OS X are available for sale at CompUSA, that restriction is going to disappear.
First off, did you read my fourth paragraph at all? Sure, you're not as abrasive as the person that says 'you're an idiot if you don't believe _____,' but what you're saying is essentially the same thing.
The goal is not to be abrasive, but sometimes a bit of abrasion is unavoidable. Yes, I think its idiotic if you think the earth was literally created in six days, just as I think it is idiotic if you think the earth is flat or if you think that Zambia will win the World Cup. I'm sorry I have to say it, but there is nothing wrong in calling a spade a spade.
Secondly, I was not intending to use the term 'Chritian scientist' to refer to the 'Christian Science' movement, but rather to a Scientist who is a Christian.
A scientist who is Christian cannot believe in the literal truth of the Bible and still claim to be thinking rationally. Believing in God is one thing --- we cannot prove nor disprove his existance so one guess is as good as the other. Believing that the earth was created in six days, despite the fact that it flies in the face of all of our knowledge, that is just silly.
Lastly, when you say "Science is predicated on a belief that the scientific method is valid," are you saying that science assumes that all things can be proven using the scientific method?
Science assumes that following the scientific method will lead to valid conclusions. This is not to say that all things are provable using the scientific method, but rather that which is provable is provable using the scientific method.
In a way, science is really no different than religious fundamentalism. It relies on the acceptance of a premise, and bases its structure on that premise. If that premise is not accepted, is not believed, the rest of it has no meaning. What differentiates the two is not one of fundemental nature, but the fact that science has time and time again validated the plausability of its premise, while the plausability of the fundamentalist position grows ever weaker.
It is simply the fact that some beliefs are more likely to be true than other beliefs. Yes, anything is possible, but certain things are more plausible. The conclusions of science are simply more plausible than the conclusions of a literal interpretation of the Bible, and that plausibility makes it a "better" belief.
That aside, let's get to the heart of the situation here. Abstract ideology is pointless to a temporal creature such as man. What makes both science and religion good is its power for bettering mankind. Science has saved countless millions of lives, and religion has made countless millions of lives better. Religious fundamentalism can do nothing to make humanity better, because it is simply ideology --- belief for the sake of belief. If tomorrow it was proved that the Bible is literally true, what does that give humanity? What use is it to know that the world was created in six days? Does that knowledge change the nature of Christ's teachings? Or is it just a fictional "fact" that would be useless even if it were true?
I would argue that it is our capability for moral action that shows we are made in God's image.
Morality is merely a check against the darker nature of man. It is simply an expression of the innate tendency of creatures to avoid behavior that is destructive to their species. Apes display the same tendencies as humans in this regard. They generally do not go about killing their own kind (no more so than humans, anyway), they will take in orphaned children, they act to preserve their group. Are apes moral, or is what passes for morality merely instinct?
For example, the invention of the atomic bomb shows that science, no matter however powerful, does not equal freedom. Instead, for fifty years we lived in fear annihilation by our own creation and by each other.
For every person killed by the atom bomb, thousands more were saved by the green revolution.
You are right that both science and religion are tools wielded by man, and man is capable of both good and evil. But my contention is not with religion in general, but "religious science" specifically. "Christian science" has no power to do good --- it is just empty ideology.
Another problem that I see is that people refuse to believe that scientific data is interpretative (i.e., it's not easy [if it's possible at all] to prove what caused _____). The Christian scientist first accepts what he believes about the Bible, and then interprets the scientific data in one way. This is not unusual, however, as the evolutionary scientist accepts what he believes from the teachings of Darwin first, and then interprets the scientific data in another way.
Science and Christianity are both predicated on belief. Christianity is predicated on a belief that the tenents of the Bible are true. Science is predicated on a belief that the scientific method is valid. From that first assumption, both Christians and scientists derive their understanding of the world.
Here is the rub: the scientists' beliefs are better. I can't really put it any other way, and I'm not willing to succumb to the popular weakness of accepting all beliefs as being equally valid. I'm an engineer --- and I see a world around me that exists because of the power of science. Science has elevated man from being little better than particularly clever animals to being something that can justifiably be said to have been made in God's image. What has "Christian science" brought humanity? Christianity as a religion can be said to have bettered the lives of many people, but "Christian science" is a mere intellectual exercise that is useful neither to the secular world nor to the religious world.
If that were the case, the G5 and the P4 wouldn't be dead ends. If that were the case, G4 wouldn't be getting more IPC than the G5. If that were the case, PIII wouldn't be getting more IPC than the P4. Pipeline does matter, in the real world.
Try to read the statement. "If the branch predictor is good enough, pipeline length doesn't matter". Pipeline length matters, but if the branch predictor is good enough, that deficit can be covered. It follows from the mathematics of the situation. A 14-stage CPU with a branch predictor that is 97% accurate actually wastes about 20% fewer cycles on mispredicted branches than a 10-stage CPU with a branch predictor that is 95% accurate. The G4 has a branch predictor comparable to the P3's. The P4's branch predictor is substantially better than the P3's, though not enough to make up for its pipeline length, and Yonah's is better-still.
The P4 gets low IPC because the branch predictor isn't good enough. The G5 gets lower IPC, despite having a relatively good branch predictor, because of a two-cycle integer latency and some poorly-optimized group formation rules. POWER5 has a 15 stage pipeline (Yonah is 12-13, Merom will be 14), and gets 800 SPECint/GHz, which is substantially better than the Alpha, which has a 9 stage pipeline.
Instruction set matters. Register file size matters.
Instruction set matters, but the particulars of modern CPUs have made things favorable for CISC as opposed to RISC. First of all, the G4 and Yonah (or to be fair, Merom, since the dual-core G4 isn't out yet, just like Merom), are both RISC inside. The price modern x86 chips pay for the x86 instruction set is a few extra decode stages. Again, with a good branch predictor, that's not a big deal. At the same time, x86 actually gains them a lot. First, its substantially denser than RISC. This effectively increases the size of the instruction cache without using extra transistors or hurting access latency. Second, it allows the expression of LOAD/STORE + OP in one instruction. Modern CPUs are very limited in their issue width, because chip area and wire delay increase quadratically with this parameter. Yonah and Opteron widen their effective issue width by fusing the internal micro-ops generated by an x86 LOAD/STORE+OP instruction, and issueing them as a single unit.
As for register file, it matters to a point. The gains for going from 8 to 16 is on the order of 10-20% for integer code. The gains for going from 16-32 are substantially less. The difference is made even more minimal considering that the Opteron and Yonah have better memory subsystems than either the G4 or G5, because of their ability to issue two memory operations per cycle (the G4 can only do one), and their ability to fuse memory operations into arithmetic operations to keep them from chewing up dispatch bandwidth.
That's comparing gcc (which isn't SPEC-optimised) against Intel's compiler, and... is that taking dual-core results and dividing by two, or is that the single-core versus the single-core? That makes a BIG difference to the multithreading benchmarks.
First, Apple's benchmarks were done using IBM's XLC (which is SPEC-optimized) and the Intel's compiler. There is probably no better compiler for each respective architecture. And taking the dual-core benchmark and dividing by two is actually the best-case scenario for the single-core chip. The real speedup from dual-core on a multithreaded benchmark will be strictly less than 2, so dividing the dual-core result by 2 to get the single core result is actually the most pessimistic way to estimate it. That't not really an issue here, because SPECint_rate is literally just two completely seperate instances of SPECint, so the speedup is very nearly 100% from dual-core.
OK, that's changed. They previously listed it as half clock. 667MHz on each of two busses is still nothing to sneeze at.
No its not, and the MPC8641D will indeed have more memory bandwidth than either Yonah or Merom. However, most integer code i
Actually, I made a math error. Cell's DP theoretical performance is the same as a dual-core Opteron's (10 gigaflops), since the Opteron has 2 FPU pipelines and 2 cores.
Cell's peak theoretical performance is 25 gigaflops, derived by taking the product of the clockspeed (3.2 GHz), and the number of operations per cycle (8). In reality, this figure is highly optimistic. Each SPE only has a single floating-point pipeline. The 8 operations/cycle figure is derived by counting a 4-element single-precision multiply-accumulate as 8 total operations. Moreover, when doing double-precision operations, it takes an additional 5x speed hit, since they must be performed in multiple clock cycles. That results in Cell's theoretical performance for double-precision code being a total of 10x lower (according to IBM), or around 2.5 gigaflops per SPE. At 10 gigaflops per chip, that's still relatively impressive, compared to the 5 gigaflops per chip a dual-core Opteron (2.4 GHz) can handle, but the actual performance of a Cell chip is going to be a lot less than the actual performance of the Opteron.
Understanding why requires a bit of understanding of chip-design, but the basics are simple. The Cell SPE basically has four things working against it:
1) No dynamic branch prediction. This means that when the Cell SPE encounters a branch instruction, it will always assume the backwards branch is taken. This works fine for loops, where its good to assume that the branch at the end of the loop will jump back to the beginning of the loop, but doesn't work well for anything else. If the guess is wrong, then the CPU pays an 18 cycle penalty while the pipeline is flushed and the correct branch path is followed. The Opteron, on the other hand, keeps track of the history of each branch. It can then make a much better guess about which way the branch will go, and avoid paying a penalty for guessing wrong. Since the Opteron's pipeline is shorter, this also means the penalty for an incorrect guess is much less (around 12 cycles). The net result of all this is that if your code has lots of short loops (static branch prediction always mispredicts the iteration that exists the loop), or a lot of complex control flow, Cell's SPE's are going to lose a lot of their theoretical performance since many cycles will be wasted on mispredicted branches.
2) Very high latency for instructions and loads. In Cell, the floating-point latency is at least six cycles, and the load latency from the local store is at least 6 cycles. For Opteron, its 4 cycles and 3 cycles, respectively. Basically, the instruction latency tells you by how many clock cycles you must seperate dependent operations. Eg: on an Opteron, you can issue a memory load, and assuming an L1 cache hit, you can issue an instruction that uses the loaded register 3 cycles later. If you have no instructions you can issue until that load is completed, then you just issue nothing that cycle and lose some of your potential throughput. Since the SPE's latencies are much higher, there is a much higher chance that you won't have any non-dependent instructions to issue on a given cycle, and must waste that cycle.
3) A very specialized memory model. Cell's SPEs can only directly address 256KB of local memory. If you have data bigger than that, you have to manually shuffle it in and out of that local memory. The latency for doing this shuffling is extremely high on Cell. This means that in code that accesses big data sets, if you can't effectively partition your data sets, you'll waste a lot of time shuffling things in and out of memory.
4) No out-of-order execution. Modern CPUs like an Opteron will rearrange your instructions to get around the instruction latencies I mentioned earlier. It'll look ahead in the code stream a couple of dozen instructions to find non-dependent ones that can be issued while waiting for other ones to finish. Cell won't do that. If you have an ADD in your code, and then right after you have a MUL that uses the results of the ADD, then Cell will merrily wait 6 cycles waiting for the ADD to finish, even if right after the MUL you have another ADD that doesn't need the results of the first one. This places a lot of burden on the c
Slashdot Mirror
The underlying foundation of XGL is exactly what a compositing window system needs --- a generic OpenGL stack that unifies control of the GPU into a single multclient-aware driver. Yes, it complicates the driver. The driver needs to properly handle and schedule multiple rendering clients, it must manage do good memory management and video memory virtualization, and it must properly handle synchronization of rendering. However, handling these things properly is The Right Solution. Aiglx just perpetuates the driver sharing lunacy that exists now, which is something that OS X and Vista will get rid off.
The reason the Mac UI works so well is that its eyecandy is a method of subtly including information that might otherwise be lost.
Your point is well-taken, but I'd suggest you sit down with a copy of OS X 10.0 sometime. The eye-candy was pretty unsubtle back then. The refinement present in Tiger took Apple several years to get right. XGL is not yet a year old. Give it some time to mature.
XGL isn't a rewrite of the server. It's a rewrite of the DDX (device-dependent) portion. That's probably the best part of the server to rewrite though, given the DIX (device-independent) is relatively clean code. XGL doesn't get rid of X's cruftiest part, though, which is xlib. XCB is ready to be a replacement, but GNOME won't be able to move to it until 3.x, because Xlib is implicitly a part of the GDK ABI.
That said, I wouldn't say XGL is better than OS X yet. OS X can do the effects you listed, it just doesn't do a lot of them for asthetic reasons. Technically, I'd argue OS X's approach is superior to XGL's, since Quartz 2D Extreme uses a direct-rendering model as opposed to XGL's indirect model. Additionally, the fact that the compositor is seperate from the window server in XGL makes synchronization a much bigger PITA than in OS X. On the other hand, the indirect model allows the X server to access the geometry stream, which allows some effects the direct-rendering model doesn't. Technical merits aside, OS X still wins because its already a stable, mature, and widely used technology. It'll be awhile before XGL is as mature as Quartz (especially at the driver layer --- DRI is really not ready for XGL yet), and before GNOME/KDE apps use vector graphics as widely as OS X apps do.
However, I agree that war sux and any civilian death is a travesty. Everyone in America, including the leadership and military, agrees. Can you say the same about the Arab world?
Frankly, nobody in the United States gives a shit about civilian death in the Middle East. Honestly, nobody here thinks twice about it. We've got our own problems, and we really don't care about theirs. Believe it or not, the Arab world is the same way. They really couldn't care less, because it doesn't affect them directly. You average Arab thinks about the death-toll resulting from terrorism about as often as your average grocery store worker thinks about the death-toll resulting from war in the Middle East.
Stop comparing the true religion of peace with that of murderous thugs.
Hmm, let's make a tally.
Deaths resulting from True Religion of Peace (TM):
Slave trade: 40 million.
French religious wars: 3 million.
Deaths resulting from Religion of Murderous Thugs (TM):
Um, you guys killed 40 million niggers? Shut man, we give up!
Regarding 3: You should be aware that radical Muslim clerics make up whatever Islamic laws they want. Their view of Islamic law has no bearing on reality. You're absolutely right, though, that a tolerable Muslim is one who does not follow the practices of his faith to the letter. Of course, the same is true of Christians. A real honest-to-God follow every letter of the Bible Christian would be quite a monstrous person based on our modern standards.
But seriously, you're joking right? If Christians were like Muslims, all those prominent gay marriage ceremonies a while ago would have been fire-bombed. Or do you think if someone tried to have a prominent, public gay marriage in, say, ANY Muslim country, they would be allowed?
It's not a Christian versus Muslim thing here. It's a rich country versus poor country thing. I'd love to see how a prominent, public gay marriage would go over in 18th century peasant Russia. No firebombs, probably, with that level of technology, but pitchforks would abound.
The real problem is not that we're dealing with Islam, but rather that we're dealing with poor and poorly educated people from culturally-backwards parts of the world. Attacking Islam won't solve anything, it's just a diversion for those who want to take action but don't want to really think about what action will be effective.
There is something serious sick in Islam today and unless some of their leaders stand up and turn the ship around,
That's really part of the problem. Islam doesn't really have any leaders. There is nobody in charge, and no formal hierarchy. It's mostly just a bunch of mobs controlled by Mullahs with a limited range of influence.
I was talking to a friend the other day, and trying to figure out a good analogy for the rarity of acts of women raping men. So I say "a women rapist is about as common as... ", and my friend fills in "a black person who doesn't commit crime?" I thought it was pretty funny, and I know he meant it entirely in jest, but I know if I ever said that out-loud in down-town DC, I'd get my ass seriously kicked. Some cartoonist, living in a place with a large and recently-emigrated (read: poor and relativey uneducated) Muslim population, depicted Mohammed as a terrorist. What did he expect? I believe its his right to say what he wishes, but some actions are just ill-advised. If I (I'm India, btw) went into the heart of the Bible Belt and started spouting off what I really thought about Christian conservatives, do you really think I wouldn't get a few death threats?
I agree with you that Muslims need to grow a thicker skin, but pinning the blame on Islam is like saying that black people are easily offended because they are black. It completely takes the cultural element out of the equation. There are a lot of people who can't take a joke, from women to hispanics to blacks to Christians to Arabs. I'd prefer to believe the reasons are cultural rather than the result of their innate tendencies...
If a religion isn't useful at getting people to love one another and not murder one another, its a failure as a religion. While you can't necessarily judge a religion based upon the actions of its followers, you would have to be an idiot to not draw some empirical evidence from these violent events about the nature of the Islam religion.
Throughout history, religion has been used as a political tool. During the wars of religion in Europe, tens of thousands of Christians slaughtered each other on the basis of Protestantism versus Catholicism. 3 million people died in France's wars of religion. Was that not evil?
The basic thing you fail to realize is that civil unrest is caused by economic, social, and political tension. The Muslim nations have more than their share of these things. The reason Christian societies have outgrown such primitive forms of violence* is because they are, on the whole, mature and economically-sound societies. At times when Muslim societies have had economic parity with Christian ones, their level of civility was quite comparable.
The evidence of history strongly suggests that religion is mostly uncorrelated with the behavior of populations. The Eastern religions are a lot mellower than any of the Semitic ones, yet that never stopped the Chinese and Japanese from having very brutal empires. Heck, Christianity did nothing to mitigate the brutality of the colonial powers either, for that matter.
*) They embrace, of course, more sophisticated and subtle, but ultimately more destructive forms of violence like launching cruise missiles from thousands of miles away...
We know for a fact that Muhammed led troops in war and had multiple wives. Seems wrong to me.
The personal character of Mohammed is something of a red-herring. Mohammed is not divine, he is simply the first leader of the faith. His place in Islam is more comparable to the place of someone like Charlemagne in Christianity rather than Jesus. As for polygamy, that's an anachronistic argument. Sects of Christianity have accepted polygamy up to the last century, and at the time when Mohammed lived, polygamy was well-accepted. I should point out that the Bible doesn't actually make any condemnations of polygamy --- that is a practice that has become part of Christianity as it has evolved.
In Christianity, the man Jesus Christ (Iesous Christos in greek) spoke at great length about how his true disciples should be harmless as doves, turn the other cheek, give, love, and forgive... They are not following the example set by Christ and therefore are committing acts of evil.
You can't just pick and choose the parts of the Bible you like. Yes, Christ said a lot of good things. But Christianity isn't merely the teachings of Christ. It includes everything in the Bible, along with the various accumualted doctrines in the case of Catholicism. There is a lot of distasteful stuff in the rest of the Bible. At the same time, there are many good things in the Quran too, along with several distasteful things.
Killing? obviously evil. no one should do it.
So our soldiers in Iraq are comitting evil acts?
Does it mean I'm gonna go killing fags?
Damn it. Here I thought I was replying to a legitimate post, now I realize I'm just talking to a half-wit.
Two things that have always bothered me about islamic martyrs: 1) unlike Jesus, and Stephen, who were peaceful martyrs killed for their words, islamic martyrs are just suicide bombers.
In theory, Islamic martyrs are
2) the koran says that muslims who martyr themselves will get however many virgins right?
Many respected scholars view that as a mistranslation, pointing out that the original text makes more sense as saying they will recieve "sweet fruits and drink" than virgins. But you will of course believe whatever you want to believe.
but in islam, women are treated as subhuman and worthless, not allowed to have rights.
Women in Islam have
AMD hasn't changed sockets just for kicks. The 754 to 939 transition was to add extra pins for the dual-channel memory controller. The AM2 socket transition will be to add support for DDR2 memory. These things required not just extra pins, but extra traces on the motherboard. Moreover, the traces have different timing characteristics because of the change in memory type. So even if AMD had used a socket with extra pins, old motherboards still wouldn't have the right lines to connect them to.
How is it a taskbar? I doesn't hold any tasks!
As for spatial browsing --- the OS X Finder is the most widely hated piece of the OS among Apple users...
What second taskbar? There is only one taskbar. The thing at the top is a menu bar. It's there because the top and bottom of the screen are easiest to access, so it makes sense to use them fully for important functions like launching and switching apps.
Now, for the confusing part. Why was their previous allocator so lame compared to malloc()?
Because glibc's malloc() is actually a pretty fast and scalable piece of code for a general-purpose memory allocator. Even GNOME's new special-purpose allocator only gets about twice the performance of glibc's.
There is really not a whole lot of benefit in making the rockets reusable. The current SRBs are solid fuel boosters, which means they are basically just metal shells filled with a propellent/oxidizer mix. The metal shells are reused, but even if they weren't the savings wouldn't be significant.
There is already a Space Shuttle in the Smithsonian. There isn't enough space for one at the DC location, so there is one at the Dulles Airport annex.
To use your analogy, it'd be like a book coming with a sticker that states "You must be able to read English to use this book!" Yes, you can buy any book you like, but you can't complain too loudly when you buy a book you can't read. The publisher doesn't have to make it available in your language.
Nobody is asking Apple to make OS X available on generic PCs. They are just saying they have no power to sue if people manage to make the OS run on generic PCs. Your analogy is thus incorrect. It's more like the book comes with a sticker saying "you must read English to read this book", and the publisher sues you if have a friend translate it into Russian so you can read it.
The GPL is not a useage contract, it is a copyright license. It governs redistribution of a work, something that is entirely within its power as a copyright license. Apple's EULA is supposedly a contract, but the legal arguments for "opt out" contracts are not exactly solid. Moreover, contract law is a large and complex subject, and it is not at all clear that Apple is allowed to set the kind of conditions they do in their EULA. If a publisher had a shrink-wrapped EULA saying that you could only keep that book on specially-approved shelves, it is very probable that suits over violations of that EULA would be thrown out of court.
That said, at the moment it is not possible to use OS X on Intel without violating Apple's copyright. Since the OS X 10.4.4 binaries are only available via purchase of a new Mac, and the Mac itself comes with a copy of the OS installed, the only way to use OS X without violating Apple's copyright by copying the binaries is to delete it from the Mac and use the restore disc on another machine. When shrink-wrapped copies of OS X are available for sale at CompUSA, that restriction is going to disappear.
First off, did you read my fourth paragraph at all? Sure, you're not as abrasive as the person that says 'you're an idiot if you don't believe _____,' but what you're saying is essentially the same thing.
The goal is not to be abrasive, but sometimes a bit of abrasion is unavoidable. Yes, I think its idiotic if you think the earth was literally created in six days, just as I think it is idiotic if you think the earth is flat or if you think that Zambia will win the World Cup. I'm sorry I have to say it, but there is nothing wrong in calling a spade a spade.
Secondly, I was not intending to use the term 'Chritian scientist' to refer to the 'Christian Science' movement, but rather to a Scientist who is a Christian.
A scientist who is Christian cannot believe in the literal truth of the Bible and still claim to be thinking rationally. Believing in God is one thing --- we cannot prove nor disprove his existance so one guess is as good as the other. Believing that the earth was created in six days, despite the fact that it flies in the face of all of our knowledge, that is just silly.
Lastly, when you say "Science is predicated on a belief that the scientific method is valid," are you saying that science assumes that all things can be proven using the scientific method?
Science assumes that following the scientific method will lead to valid conclusions. This is not to say that all things are provable using the scientific method, but rather that which is provable is provable using the scientific method.
In a way, science is really no different than religious fundamentalism. It relies on the acceptance of a premise, and bases its structure on that premise. If that premise is not accepted, is not believed, the rest of it has no meaning. What differentiates the two is not one of fundemental nature, but the fact that science has time and time again validated the plausability of its premise, while the plausability of the fundamentalist position grows ever weaker.
It is simply the fact that some beliefs are more likely to be true than other beliefs. Yes, anything is possible, but certain things are more plausible. The conclusions of science are simply more plausible than the conclusions of a literal interpretation of the Bible, and that plausibility makes it a "better" belief.
That aside, let's get to the heart of the situation here. Abstract ideology is pointless to a temporal creature such as man. What makes both science and religion good is its power for bettering mankind. Science has saved countless millions of lives, and religion has made countless millions of lives better. Religious fundamentalism can do nothing to make humanity better, because it is simply ideology --- belief for the sake of belief. If tomorrow it was proved that the Bible is literally true, what does that give humanity? What use is it to know that the world was created in six days? Does that knowledge change the nature of Christ's teachings? Or is it just a fictional "fact" that would be useless even if it were true?
I would argue that it is our capability for moral action that shows we are made in God's image.
Morality is merely a check against the darker nature of man. It is simply an expression of the innate tendency of creatures to avoid behavior that is destructive to their species. Apes display the same tendencies as humans in this regard. They generally do not go about killing their own kind (no more so than humans, anyway), they will take in orphaned children, they act to preserve their group. Are apes moral, or is what passes for morality merely instinct?
For example, the invention of the atomic bomb shows that science, no matter however powerful, does not equal freedom. Instead, for fifty years we lived in fear annihilation by our own creation and by each other.
For every person killed by the atom bomb, thousands more were saved by the green revolution.
You are right that both science and religion are tools wielded by man, and man is capable of both good and evil. But my contention is not with religion in general, but "religious science" specifically. "Christian science" has no power to do good --- it is just empty ideology.
Another problem that I see is that people refuse to believe that scientific data is interpretative (i.e., it's not easy [if it's possible at all] to prove what caused _____). The Christian scientist first accepts what he believes about the Bible, and then interprets the scientific data in one way. This is not unusual, however, as the evolutionary scientist accepts what he believes from the teachings of Darwin first, and then interprets the scientific data in another way.
Science and Christianity are both predicated on belief. Christianity is predicated on a belief that the tenents of the Bible are true. Science is predicated on a belief that the scientific method is valid. From that first assumption, both Christians and scientists derive their understanding of the world.
Here is the rub: the scientists' beliefs are better. I can't really put it any other way, and I'm not willing to succumb to the popular weakness of accepting all beliefs as being equally valid. I'm an engineer --- and I see a world around me that exists because of the power of science. Science has elevated man from being little better than particularly clever animals to being something that can justifiably be said to have been made in God's image. What has "Christian science" brought humanity? Christianity as a religion can be said to have bettered the lives of many people, but "Christian science" is a mere intellectual exercise that is useful neither to the secular world nor to the religious world.
The mass of the fuel increases, but the energy contained in its chemical bonds does not.
If that were the case, the G5 and the P4 wouldn't be dead ends. If that were the case, G4 wouldn't be getting more IPC than the G5. If that were the case, PIII wouldn't be getting more IPC than the P4. Pipeline does matter, in the real world.
Try to read the statement. "If the branch predictor is good enough, pipeline length doesn't matter". Pipeline length matters, but if the branch predictor is good enough, that deficit can be covered. It follows from the mathematics of the situation. A 14-stage CPU with a branch predictor that is 97% accurate actually wastes about 20% fewer cycles on mispredicted branches than a 10-stage CPU with a branch predictor that is 95% accurate. The G4 has a branch predictor comparable to the P3's. The P4's branch predictor is substantially better than the P3's, though not enough to make up for its pipeline length, and Yonah's is better-still.
The P4 gets low IPC because the branch predictor isn't good enough. The G5 gets lower IPC, despite having a relatively good branch predictor, because of a two-cycle integer latency and some poorly-optimized group formation rules. POWER5 has a 15 stage pipeline (Yonah is 12-13, Merom will be 14), and gets 800 SPECint/GHz, which is substantially better than the Alpha, which has a 9 stage pipeline.
Instruction set matters. Register file size matters.
Instruction set matters, but the particulars of modern CPUs have made things favorable for CISC as opposed to RISC. First of all, the G4 and Yonah (or to be fair, Merom, since the dual-core G4 isn't out yet, just like Merom), are both RISC inside. The price modern x86 chips pay for the x86 instruction set is a few extra decode stages. Again, with a good branch predictor, that's not a big deal. At the same time, x86 actually gains them a lot. First, its substantially denser than RISC. This effectively increases the size of the instruction cache without using extra transistors or hurting access latency. Second, it allows the expression of LOAD/STORE + OP in one instruction. Modern CPUs are very limited in their issue width, because chip area and wire delay increase quadratically with this parameter. Yonah and Opteron widen their effective issue width by fusing the internal micro-ops generated by an x86 LOAD/STORE+OP instruction, and issueing them as a single unit.
As for register file, it matters to a point. The gains for going from 8 to 16 is on the order of 10-20% for integer code. The gains for going from 16-32 are substantially less. The difference is made even more minimal considering that the Opteron and Yonah have better memory subsystems than either the G4 or G5, because of their ability to issue two memory operations per cycle (the G4 can only do one), and their ability to fuse memory operations into arithmetic operations to keep them from chewing up dispatch bandwidth.
That's comparing gcc (which isn't SPEC-optimised) against Intel's compiler, and... is that taking dual-core results and dividing by two, or is that the single-core versus the single-core? That makes a BIG difference to the multithreading benchmarks.
First, Apple's benchmarks were done using IBM's XLC (which is SPEC-optimized) and the Intel's compiler. There is probably no better compiler for each respective architecture. And taking the dual-core benchmark and dividing by two is actually the best-case scenario for the single-core chip. The real speedup from dual-core on a multithreaded benchmark will be strictly less than 2, so dividing the dual-core result by 2 to get the single core result is actually the most pessimistic way to estimate it. That't not really an issue here, because SPECint_rate is literally just two completely seperate instances of SPECint, so the speedup is very nearly 100% from dual-core.
OK, that's changed. They previously listed it as half clock. 667MHz on each of two busses is still nothing to sneeze at.
No its not, and the MPC8641D will indeed have more memory bandwidth than either Yonah or Merom. However, most integer code i
Actually, I made a math error. Cell's DP theoretical performance is the same as a dual-core Opteron's (10 gigaflops), since the Opteron has 2 FPU pipelines and 2 cores.
Cell's peak theoretical performance is 25 gigaflops, derived by taking the product of the clockspeed (3.2 GHz), and the number of operations per cycle (8). In reality, this figure is highly optimistic. Each SPE only has a single floating-point pipeline. The 8 operations/cycle figure is derived by counting a 4-element single-precision multiply-accumulate as 8 total operations. Moreover, when doing double-precision operations, it takes an additional 5x speed hit, since they must be performed in multiple clock cycles. That results in Cell's theoretical performance for double-precision code being a total of 10x lower (according to IBM), or around 2.5 gigaflops per SPE. At 10 gigaflops per chip, that's still relatively impressive, compared to the 5 gigaflops per chip a dual-core Opteron (2.4 GHz) can handle, but the actual performance of a Cell chip is going to be a lot less than the actual performance of the Opteron.
Understanding why requires a bit of understanding of chip-design, but the basics are simple. The Cell SPE basically has four things working against it:
1) No dynamic branch prediction. This means that when the Cell SPE encounters a branch instruction, it will always assume the backwards branch is taken. This works fine for loops, where its good to assume that the branch at the end of the loop will jump back to the beginning of the loop, but doesn't work well for anything else. If the guess is wrong, then the CPU pays an 18 cycle penalty while the pipeline is flushed and the correct branch path is followed. The Opteron, on the other hand, keeps track of the history of each branch. It can then make a much better guess about which way the branch will go, and avoid paying a penalty for guessing wrong. Since the Opteron's pipeline is shorter, this also means the penalty for an incorrect guess is much less (around 12 cycles). The net result of all this is that if your code has lots of short loops (static branch prediction always mispredicts the iteration that exists the loop), or a lot of complex control flow, Cell's SPE's are going to lose a lot of their theoretical performance since many cycles will be wasted on mispredicted branches.
2) Very high latency for instructions and loads. In Cell, the floating-point latency is at least six cycles, and the load latency from the local store is at least 6 cycles. For Opteron, its 4 cycles and 3 cycles, respectively. Basically, the instruction latency tells you by how many clock cycles you must seperate dependent operations. Eg: on an Opteron, you can issue a memory load, and assuming an L1 cache hit, you can issue an instruction that uses the loaded register 3 cycles later. If you have no instructions you can issue until that load is completed, then you just issue nothing that cycle and lose some of your potential throughput. Since the SPE's latencies are much higher, there is a much higher chance that you won't have any non-dependent instructions to issue on a given cycle, and must waste that cycle.
3) A very specialized memory model. Cell's SPEs can only directly address 256KB of local memory. If you have data bigger than that, you have to manually shuffle it in and out of that local memory. The latency for doing this shuffling is extremely high on Cell. This means that in code that accesses big data sets, if you can't effectively partition your data sets, you'll waste a lot of time shuffling things in and out of memory.
4) No out-of-order execution. Modern CPUs like an Opteron will rearrange your instructions to get around the instruction latencies I mentioned earlier. It'll look ahead in the code stream a couple of dozen instructions to find non-dependent ones that can be issued while waiting for other ones to finish. Cell won't do that. If you have an ADD in your code, and then right after you have a MUL that uses the results of the ADD, then Cell will merrily wait 6 cycles waiting for the ADD to finish, even if right after the MUL you have another ADD that doesn't need the results of the first one. This places a lot of burden on the c