Since I've seen posts from at least one kernel developer in response to the attached story, I figured that this might be a good place to ask the following question:
A little while ago, I wrote an application that uses an incredible amount of memory... A very space inefficient implementation of
Eratosthenes' Sieve. In essence, what the algorithm does is cycle through the entire contents of memory sequentially many, many times (not a completely correct description). What I found with the following three kernel versions:
2.4.4
2.4.8
2.4.17
is that any time the program's footprint exceeded the physical ammount of available memory, performance degraded exponentially. I found this to be very suprising, considering that I was only exceeding the physical 1024 Megs of memory by less than 10 Megs. About the only difference between the three kernels I listed above is that the 2.4.17 kernel would kill of memory intensive processes a lot quicker than the other 2 versions.
My question for the Kernel gods out there is as follows: are there any stable 2.4.x kernel releases out there that would handle this type of stress without the performance degradation that I've experienced with these kernels?
If you think about it, it's not that suprising that a compiler/assembler written by the chip manufacturer generates faster binaries. Having access to the chip specs and the engineers that designed the chip virtually ensures that this will indeed be the case.
Before people start to predict gcc's demise, I would like to ask the reviewers if they tried benchmarking the performance of the intel compiler's binaries on a Transmeta or AMD chip. I think that might lead to us seeing the Intel compiler in a slightly different light.
It's good to here some good news come from SGI... I've been a fan of pretty much everything that they've ever done. I find Irix to have the sweetest desktop out there of any Unixes I've ever used (Gnome and KDE pundits may repectfully disagree). Hell, even the cases they put their machines in are works of art.
Anyway, although I am rather happy to hear that they will be recieving a financial shot in the arm from the new wave of government spending, I am a little worried, given the track record for stability of Irix, that these machines will be running fighter jets (can anyone say kernel panic at 30,000 feet)?
I work in the ITV industry and I have to say that this will not happen overnight. Most digital cable plants out there broadcast over a 27 Mbit pipe. This pipe is not wide enough to accomodate more than a couple of channels of really low quality VOD (they lower the quality of the mpeg compression to accomodate the bandwidth).
In order for most cable plants to offer true high quality video on demand with more selections than this, they have to upgrade everything from the equipment in the cable plant to the wire running into your home. Given the speed at which Cable Companies change the technology that they use, I give this five to ten years before we actually see it.
Actually, Mc Affee and other AV products detect things other than viruses... I.E. the Code Red Worm. Please make sure you understand what the difference between a worm, virus, and trojan horse are before you post a bogus definition in the future.
Oh, and by the way, I believe this magic latern horseshit is actually a trojan horse, not a virus or a worm.
I guess this is what you get when a bunch of web monkeys try to be encyclopedists...
Pardon me for suggesting something a little over the top, but why not use latex2pdf to generate your reports in pdf format... Using the cgi language of your choice, you can generate the LaTeX source and then call latex2pdf to generate the postscript...
While not the simplest solution to your problem, it is industrial strength, LaTeX provides any and all formatting capabilities that you would ever need (provided you invest the time to learn:) and PDF is a (almost) open document format.
Being a college graduate (and having BS's in both math and c.s.), I've experienced both really good and really bad teachers in my fields of study... The following is my insight on what make excellent and awful teachers in the aforementioned fields:
Good teachers are the ones who have some type of plan of what they want their students to learn during the course of a class... Who really want to present their students with the knowledge and skills that they see as important. They then devise useful exercises and projects that not only reinforce the class material, but challenge the students to develop a deeper understanding of the material and to start to apply the information in new and creative ways.
For lack of a better example, I remember my teacher for the first class I took in C/C++ had us write a string processing program that first used char arrays as the basic data structure to process, and then, as our next assignment, had us rewrite the program to char pointers... It doesn't seem like a very useful thing to have done, but it reinforced for us that arrays and pointers in C are virtually the same thing. (I know language lawyers will tend to disagree).
Moving on to my perception of what a bad teacher is, I have also had an incredible number of these... They are the ones that either make the material to learn incomprehensible to the average person in the class, or move through a given (usually very small) set of sections of a textbook at a pace that serves no one except their own laziness. These teachers have no goal as to what they want their students to learn and are teaching for no other reason than they cannot hold any other job.
First, let me say that I think what Metallica's doing is truly absurd, but let's look at who's truly breaking the law...
Is it a university b/c they are acting as an ISP?
Is it Napster (or for that matter Sun or even microsoft) for providing a file sharing protocol?
Is it Metallica's for wanting to make money off of their efforts?
No! it's the users... people using napster are stealing.
This is not a battle between napster and metallica of metallica and harvard... It is instead a battle that we are fighting against the RIAA... We are outraged that a CD costs 15-20 dollars. The only solution to this is boycott any artist that is on a label associated with the RIAA.
Let me first say that I work for a small security company and have had to port things to the DIICOE environment; working with DIICOE in general is not anything that I would wish upon my worst enemy.
The DIICOE certification is not like POSIX or any other certification driven by a standards organization... You get DIICOE certification by playing the politics game (The reason that HP-UX and NT are certified).
Also, in order for a unix system to be considered DIICOE compliant, it must run a bastardized version of CDE (you know, the abortion of a Motif look-alike) as well as have a herd of memory sucking daemons running behind the scenes.
Slashdot Mirror
Since I've seen posts from at least one kernel developer in response to the attached story, I figured that this might be a good place to ask the following question:
A little while ago, I wrote an application that uses an incredible amount of memory... A very space inefficient implementation of Eratosthenes' Sieve. In essence, what the algorithm does is cycle through the entire contents of memory sequentially many, many times (not a completely correct description). What I found with the following three kernel versions:
- 2.4.4
- 2.4.8
- 2.4.17
is that any time the program's footprint exceeded the physical ammount of available memory, performance degraded exponentially. I found this to be very suprising, considering that I was only exceeding the physical 1024 Megs of memory by less than 10 Megs. About the only difference between the three kernels I listed above is that the 2.4.17 kernel would kill of memory intensive processes a lot quicker than the other 2 versions.My question for the Kernel gods out there is as follows: are there any stable 2.4.x kernel releases out there that would handle this type of stress without the performance degradation that I've experienced with these kernels?
If you think about it, it's not that suprising that a compiler/assembler written by the chip manufacturer generates faster binaries. Having access to the chip specs and the engineers that designed the chip virtually ensures that this will indeed be the case.
Before people start to predict gcc's demise, I would like to ask the reviewers if they tried benchmarking the performance of the intel compiler's binaries on a Transmeta or AMD chip. I think that might lead to us seeing the Intel compiler in a slightly different light.
It's good to here some good news come from SGI... I've been a fan of pretty much everything that they've ever done. I find Irix to have the sweetest desktop out there of any Unixes I've ever used (Gnome and KDE pundits may repectfully disagree). Hell, even the cases they put their machines in are works of art.
Anyway, although I am rather happy to hear that they will be recieving a financial shot in the arm from the new wave of government spending, I am a little worried, given the track record for stability of Irix, that these machines will be running fighter jets (can anyone say kernel panic at 30,000 feet)?
I work in the ITV industry and I have to say that this will not happen overnight. Most digital cable plants out there broadcast over a 27 Mbit pipe. This pipe is not wide enough to accomodate more than a couple of channels of really low quality VOD (they lower the quality of the mpeg compression to accomodate the bandwidth).
In order for most cable plants to offer true high quality video on demand with more selections than this, they have to upgrade everything from the equipment in the cable plant to the wire running into your home. Given the speed at which Cable Companies change the technology that they use, I give this five to ten years before we actually see it.
Sorry to burst everyone's bubble.
Oh, and by the way, I believe this magic latern horseshit is actually a trojan horse, not a virus or a worm.
I guess this is what you get when a bunch of web monkeys try to be encyclopedists...
What happened to the text of the article:
Her speach dealt both with her love of money and
her desire to roll naked in it
Just curious...
While not the simplest solution to your problem, it is industrial strength, LaTeX provides any and all formatting capabilities that you would ever need (provided you invest the time to learn :) and PDF is a (almost) open document format.
Good teachers are the ones who have some type of plan of what they want their students to learn during the course of a class... Who really want to present their students with the knowledge and skills that they see as important. They then devise useful exercises and projects that not only reinforce the class material, but challenge the students to develop a deeper understanding of the material and to start to apply the information in new and creative ways.
For lack of a better example, I remember my teacher for the first class I took in C/C++ had us write a string processing program that first used char arrays as the basic data structure to process, and then, as our next assignment, had us rewrite the program to char pointers... It doesn't seem like a very useful thing to have done, but it reinforced for us that arrays and pointers in C are virtually the same thing. (I know language lawyers will tend to disagree).
Moving on to my perception of what a bad teacher is, I have also had an incredible number of these... They are the ones that either make the material to learn incomprehensible to the average person in the class, or move through a given (usually very small) set of sections of a textbook at a pace that serves no one except their own laziness. These teachers have no goal as to what they want their students to learn and are teaching for no other reason than they cannot hold any other job.
Is it a university b/c they are acting as an ISP?
Is it Napster (or for that matter Sun or even microsoft) for providing a file sharing protocol?
Is it Metallica's for wanting to make money off of their efforts?
No! it's the users... people using napster are stealing.
This is not a battle between napster and metallica of metallica and harvard... It is instead a battle that we are fighting against the RIAA... We are outraged that a CD costs 15-20 dollars. The only solution to this is boycott any artist that is on a label associated with the RIAA.
Let me first say that I work for a small security company and have had to port things to the DIICOE environment; working with DIICOE in general is not anything that I would wish upon my worst enemy. The DIICOE certification is not like POSIX or any other certification driven by a standards organization... You get DIICOE certification by playing the politics game (The reason that HP-UX and NT are certified). Also, in order for a unix system to be considered DIICOE compliant, it must run a bastardized version of CDE (you know, the abortion of a Motif look-alike) as well as have a herd of memory sucking daemons running behind the scenes.