Sorry but OS/application design can also help very much: BeOS was much more responsive than Linux and Windows are today on much faster CPUs (granted disk have not improved much).
Probably because of good design (~14s to boot in BeOS in a usable GUI, while Linux/Windows takes >1min) and heavy usage of multi-threading: it felt much faster.
What I find interesting is that the vector processor are restricted to single precision floating point calculations. This isn't terribly useful for scientific computations (there is the same problem with the GPU): currently the IEEE is working on a standard for 128bit precision floating point calculations!
Of course for 3D, video and sound, 32bit precision is good enough and *if* programmers (a big if) manage to overcome the pain of 'parallel programming' then it could be a big success.
Sigh, I wouldn't call this a discussion, because for a discussion to occur it must happen between people in 'good faith' and when you talk about "Richard Dawkins" (who is he? Don't know and don't care: atheist needs no priest, popes and the likes) as an "atheist priest" shows that you're in bad faith..
I don't beleive in father christmas too, does-it make me a 'christmas religious' too?
> France tried it [cut] go read the dismal record if you want to get depressed
I'm not sure where you read this 'dismal record' but as a french, I don't know about it: the laicity of state is quite well accepted, and a big percentage of french are atheist/agnostic and our society works ok.
And as for separation from church and state in the US, given that your president swear on the bible, I'd say that it is pretty shallow. Also maybe separation from church and state was well-accepted in the US but in France, it had to be fought against religious people in some occasions (some supporters of the pope took the arms for example).
And saying that atheism is a religion is a way religious people have to slander atheist, but atheists have no priest, no prayer, no mythology about the beginning or the end of the universe, no mythology about the 'after-life', something common to nearly all the religion.. So it isn't a religion, it is a *belief* ok but what makes a religion is all the other things I've listed.
Frankly your post is crap, I wonder how it was moderated so high.
But given the times it takes to develop the Hurd, this also assumes that Moore's "law" is still valid by the time, the Hurd/L4 runs the thing you need..
Jokes aside, I've heard that L4 wasn't suited very well to run on x86-64 due to a lack of segment in x86-64, is-this true?
I disagree, assembly language is even more difficult to use than C language and the programs are usually unreadable. Of course it is good for learning how things really work, but that's all.
As for OOP, it is really hard to grasp at the beginning and how to use correctly, so it should be taught. Who said anything about learning class libraries? This has nothing to do about learning how OOP works and why it is useful.
Also as for "hard science", why should school be limited to hard science?
I've learned about automated proof at school and don't expect it to become useful anytime soon (the programs conforms to the spec, great now is the spec correct?) except in very limitated situation, so I consider it like time mostly wasted.
Depends C is a good language for learning how stuff works due to its low level.
But as you noted C has also lots of pitfall in its use, which can put off many student..
Also C is a terrible language for learning high level stuff, such as object orientation, when teaching it, you don't want that the student focus on how it is implemented, but on why it is useful and how to use it..
Maybe it's just cost 1c less to disable them in software than in HW? OK, the number of users which disable the protection will be much higher, but I suspect than this number is so small that it's "lost in the noise". Also this number can't be measured while the cost saving by disabling a functionnality through software vs pushing the chip an additional tool to disable some pins is very easy to compute.
Except of course that his OS is not only running of Intel computers..
As for reimplementing Unix, you're right that it isn't particulary original, which is a *good thing* because last I looked each time an OS tries to be original it fails..
EROS failed: they never got it off the ground (I wish better luck to its successor but the 'yet another language' will need serious marketing to be accepted..), Hurd has not failed yet, but whether it will success in gathering users is hard to say, do you see a pattern here?
Beside "normal" micro-kernels such as Chorus, QNX, etc.. the only "good" original "OS" I can think of is L4, but if we measure success by the number of users..
> Hell, ATI had a card in which you could unlock four pipelines with a small program.
But remember that those four pipelines may work *or not* as it is also an easy way to disable faulty units, thus allowing the HW makers to sell at reduced price something which should have been trashed otherwise..
> - do large-scale matrix math, or other parallelizable calculations?
Because usually scientific computations are made with 64bit floating point (or above: Intel FPU provides 80bit floating points precision), 32 bit won't cut it..
Oh and don't believe the hype: when Nvidia made noise about 128bit computation, it was 4*32 which is normal in graphism (ARGB) but very different from 128 FP math..
It is finite of course but it isn't fixed: it increase each time someone creates a Word document, an IE only webpage, a HW device which works only with Windows, etc.. And it decrease each time someone use open standards or use MacOS X..
I disagree: RISC means *Reduced* Instruction Set Computer, so it means clearly that to it removes some complexity of the instruction set.
And RISC CPU like G5 or Sparc have a reduced instruction set: load/store architecture, less number of addressing mode, etc..
It is true that under the IS, the engine has become more complex (OOO execution, etc.) but the instruction set is still less complex that CPU existing before..
Each new CPU instruction after the RISC invention set is RISC, so saying that RISC is dead is wrong.
But it is true, that *on the desktop* (and nowhere else) compatibility has won over the performance increase gained by the simpler IS of RISC CPUs.
You are not totally right: even if your program is compiled in a binary using sequential intructions, the CPU usually can parallelise execution with additional HW (out of order execution, completion,etc..).
But it is true that within a sequential program, the available paralelism is very low, so I'd say that the author of the article is "wildly overoptimistic" over the performance of a cell CPU.
How vector units could help emulating x86 instructions (except for SSE of course) ????
I wouldn't call him an idiot for this article, but a 'believer': his "faith" in the cell is exagerated!!
Specialised computer are great with specialised SW but suck at general purpose computing or at emulating other (different) architecture..
OK, I'm curious: do Word, Excel, IE support AppleScript?
You know, KDE applications are scriptable too, but it' true that very few Linux users use only KDE or only Gnome apps usually it is a mix of toolkit (bleah)..
While I agree with you with the rape section, I disagree with the 'fisher price' look: when I installed WindowsXP, the first thing is that it looked 'fisher price' to me (and horrible: thankfuly you can customise it).
And I'm French and do not have kids! Many people I know had the same impression: some like it, some not. Agreed that this is not hard stat, it depends on the way it is written in the wikipedia entry..
Very few people has a 10k RPM disk as they are very expensive and there not going down in price. And speed&cachesize within the disk has not really increased over the year for reasonably priced disk: still at 7200RPM and 8MB cache since a long time, the only thing that change is capacity (access time is also reduced year after year, even with a constant RPM but the improvement is quite slow).
As for the video benchmark, it was a benchmark made of games of course and it showed that the gains where really small, but sorry I can't find the URL.
I disagree that these so fast bust are so useful: -Using ever faster IDE/ATA bus means few improvement when disk access time is not reduced at the same rate.. -I remember that a review showed going from an AGP*2 to a PCI express give very little improvement (except for low-end solution with shared memory system).
Increasing bus bandwith alone is nearly meaningless if it isn't the bottleneck in the fist place!
Sorry but I disagree with you: while Einstein contribution on SR and QM were 'in the air' in 1907, GR is really his masterpiece and he made it *alone*!
Except from Grossman teaching him the needed mathmatics for GR, but we're talking about physics here.
He was such in advance with GR that other physicist didn't understood/use his results for years!
Too bad Alain Aspect's experiment was so late!
on
100 Years of Einstein
·
· Score: 1
I'd really wish that Alain Aspect's experiment could have been made when Einstein was young..
Maybe he could have shed some light on it: think about what this experiment reveals: one interaction at one place can have an 'impact' instantly on a different location! But this 'impact' is subtle enough that it cannot provide information faster than the speed of light..
That's just plain weird! And Einstein was one of the few men who viewed how bizarre QM is and thought it was incomplete as a result: while some thought that he was 'an old fart' which cannot accept new theory, for me this is exactly the opposite: I bet that few takers of QM really understood how weird it is when you think about EPR paradox..
I still don't understand it: non-local instant interaction but which still cannot convey information faster than the speed of light???
Uh? I wonder if string theory helps here to describe a "clearer" picture of these weird distant interactions..
Of course in the asynchronous situation you can avoid having to handle overflow *if* you have very precise information on the normal situation and *if* everything goes according to the plan..
Now, I'm not against message passing IPC, but there just a different API to solve a complex problem, the API may be a little bit better, but the problem is still complex, so the improvement is not so big..
If memory serves in QNX, the message passing API is synchronous: it is simpler to manage for the kernel, bufferisation (asynchronous) is put on a layer 'on top' of the synchronous messaging API (I read it on the Internet so beware).
Slashdot Mirror
Sorry but OS/application design can also help very much: BeOS was much more responsive than Linux and Windows are today on much faster CPUs (granted disk have not improved much).
Probably because of good design (~14s to boot in BeOS in a usable GUI, while Linux/Windows takes >1min) and heavy usage of multi-threading: it felt much faster.
What I find interesting is that the vector processor are restricted to single precision floating point calculations.
This isn't terribly useful for scientific computations (there is the same problem with the GPU): currently the IEEE is working on a standard for 128bit precision floating point calculations!
Of course for 3D, video and sound, 32bit precision is good enough and *if* programmers (a big if) manage to overcome the pain of 'parallel programming' then it could be a big success.
Sigh, I wouldn't call this a discussion, because for a discussion to occur it must happen between people in 'good faith' and when you talk about "Richard Dawkins" (who is he? Don't know and don't care: atheist needs no priest, popes and the likes) as an "atheist priest" shows that you're in bad faith..
I don't beleive in father christmas too, does-it make me a 'christmas religious' too?
> France tried it [cut] go read the dismal record if you want to get depressed
I'm not sure where you read this 'dismal record' but as a french, I don't know about it: the laicity of state is quite well accepted, and a big percentage of french are atheist/agnostic and our society works ok.
And as for separation from church and state in the US, given that your president swear on the bible, I'd say that it is pretty shallow.
Also maybe separation from church and state was well-accepted in the US but in France, it had to be fought against religious people in some occasions (some supporters of the pope took the arms for example).
And saying that atheism is a religion is a way religious people have to slander atheist, but atheists have no priest, no prayer, no mythology about the beginning or the end of the universe, no mythology about the 'after-life', something common to nearly all the religion..
So it isn't a religion, it is a *belief* ok but what makes a religion is all the other things I've listed.
Frankly your post is crap, I wonder how it was moderated so high.
But given the times it takes to develop the Hurd, this also assumes that Moore's "law" is still valid by the time, the Hurd/L4 runs the thing you need..
Jokes aside, I've heard that L4 wasn't suited very well to run on x86-64 due to a lack of segment in x86-64, is-this true?
I disagree, assembly language is even more difficult to use than C language and the programs are usually unreadable.
Of course it is good for learning how things really work, but that's all.
As for OOP, it is really hard to grasp at the beginning and how to use correctly, so it should be taught.
Who said anything about learning class libraries? This has nothing to do about learning how OOP works and why it is useful.
Also as for "hard science", why should school be limited to hard science?
I've learned about automated proof at school and don't expect it to become useful anytime soon (the programs conforms to the spec, great now is the spec correct?) except in very limitated situation, so I consider it like time mostly wasted.
Depends C is a good language for learning how stuff works due to its low level.
But as you noted C has also lots of pitfall in its use, which can put off many student..
Also C is a terrible language for learning high level stuff, such as object orientation, when teaching it, you don't want that the student focus on how it is implemented, but on why it is useful and how to use it..
Well, even if some rules had some rational causes, the real question is: why still respecting deprecated rules?
Pork is not dangerous by any means currently!
> except maybe amiga
The OS without memory protection?
Great!
> As someone else already said, "people seem to forget what the R in RAM stands for".
> What kills RAM nowadays in common scenarios is latency.
No problem, to reduce latency RAM makers should just increase the speed of light..
Easy!
Maybe it's just cost 1c less to disable them in software than in HW?
OK, the number of users which disable the protection will be much higher, but I suspect than this number is so small that it's "lost in the noise".
Also this number can't be measured while the cost saving by disabling a functionnality through software vs pushing the chip an additional tool to disable some pins is very easy to compute.
Except of course that his OS is not only running of Intel computers..
As for reimplementing Unix, you're right that it isn't particulary original, which is a *good thing* because last I looked each time an OS tries to be original it fails..
EROS failed: they never got it off the ground (I wish better luck to its successor but the 'yet another language' will need serious marketing to be accepted..), Hurd has not failed yet, but whether it will success in gathering users is hard to say, do you see a pattern here?
Beside "normal" micro-kernels such as Chorus, QNX, etc.. the only "good" original "OS" I can think of is L4, but if we measure success by the number of users..
> Hell, ATI had a card in which you could unlock four pipelines with a small program.
But remember that those four pipelines may work *or not* as it is also an easy way to disable faulty units, thus allowing the HW makers to sell at reduced price something which should have been trashed otherwise..
> - do large-scale matrix math, or other parallelizable calculations?
Because usually scientific computations are made with 64bit floating point (or above: Intel FPU provides 80bit floating points precision), 32 bit won't cut it..
Oh and don't believe the hype: when Nvidia made noise about 128bit computation, it was 4*32 which is normal in graphism (ARGB) but very different from 128 FP math..
OK, your version is better than his Perl version, but it's still far less readable than his Ruby's version..
A lot of inertia, and if it is finite agreed.
It is finite of course but it isn't fixed: it increase each time someone creates a Word document, an IE only webpage, a HW device which works only with Windows, etc..
And it decrease each time someone use open standards or use MacOS X..
I disagree: RISC means *Reduced* Instruction Set Computer, so it means clearly that to it removes some complexity of the instruction set.
And RISC CPU like G5 or Sparc have a reduced instruction set: load/store architecture, less number of addressing mode, etc..
It is true that under the IS, the engine has become more complex (OOO execution, etc.) but the instruction set is still less complex that CPU existing before..
Each new CPU instruction after the RISC invention set is RISC, so saying that RISC is dead is wrong.
But it is true, that *on the desktop* (and nowhere else) compatibility has won over the performance increase gained by the simpler IS of RISC CPUs.
You are not totally right: even if your program is compiled in a binary using sequential intructions, the CPU usually can parallelise execution with additional HW (out of order execution, completion,etc..).
But it is true that within a sequential program, the available paralelism is very low, so I'd say that the author of the article is "wildly overoptimistic" over the performance of a cell CPU.
How vector units could help emulating x86 instructions (except for SSE of course) ????
I wouldn't call him an idiot for this article, but a 'believer': his "faith" in the cell is exagerated!!
Specialised computer are great with specialised SW but suck at general purpose computing or at emulating other (different) architecture..
OK, I'm curious: do Word, Excel, IE support AppleScript?
You know, KDE applications are scriptable too, but it' true that very few Linux users use only KDE or only Gnome apps usually it is a mix of toolkit (bleah)..
While I agree with you with the rape section, I disagree with the 'fisher price' look: when I installed WindowsXP, the first thing is that it looked 'fisher price' to me (and horrible: thankfuly you can customise it).
And I'm French and do not have kids!
Many people I know had the same impression: some like it, some not.
Agreed that this is not hard stat, it depends on the way it is written in the wikipedia entry..
Very few people has a 10k RPM disk as they are very expensive and there not going down in price. And speed&cachesize within the disk has not really increased over the year for reasonably priced disk: still at 7200RPM and 8MB cache since a long time, the only thing that change is capacity (access time is also reduced year after year, even with a constant RPM but the improvement is quite slow).
As for the video benchmark, it was a benchmark made of games of course and it showed that the gains where really small, but sorry I can't find the URL.
I disagree that these so fast bust are so useful:
-Using ever faster IDE/ATA bus means few improvement when disk access time is not reduced at the same rate..
-I remember that a review showed going from an AGP*2 to a PCI express give very little improvement (except for low-end solution with shared memory system).
Increasing bus bandwith alone is nearly meaningless if it isn't the bottleneck in the fist place!
Sorry but I disagree with you: while Einstein contribution on SR and QM were 'in the air' in 1907, GR is really his masterpiece and he made it *alone*!
Except from Grossman teaching him the needed mathmatics for GR, but we're talking about physics here.
He was such in advance with GR that other physicist didn't understood/use his results for years!
I'd really wish that Alain Aspect's experiment could have been made when Einstein was young..
..
Maybe he could have shed some light on it: think about what this experiment reveals: one interaction at one place can have an 'impact' instantly on a different location!
But this 'impact' is subtle enough that it cannot provide information faster than the speed of light..
That's just plain weird! And Einstein was one of the few men who viewed how bizarre QM is and thought it was incomplete as a result: while some thought that he was 'an old fart' which cannot accept new theory, for me this is exactly the opposite: I bet that few takers of QM really understood how weird it is when you think about EPR paradox..
I still don't understand it: non-local instant interaction but which still cannot convey information faster than the speed of light???
Uh?
I wonder if string theory helps here to describe a "clearer" picture of these weird distant interactions
Of course in the asynchronous situation you can avoid having to handle overflow *if* you have very precise information on the normal situation and *if* everything goes according to the plan..
Now, I'm not against message passing IPC, but there just a different API to solve a complex problem, the API may be a little bit better, but the problem is still complex, so the improvement is not so big..
If memory serves in QNX, the message passing API is synchronous: it is simpler to manage for the kernel, bufferisation (asynchronous) is put on a layer 'on top' of the synchronous messaging API (I read it on the Internet so beware).