C library for numerical analysis and math
This is my second email. will it reach? I've written a fast C library for numerical analysis. -integration -differential equations -root finding -interpolation -fourier analysis -ploting curves -misc functions such as rect, tri etc I need to know where to go from here please answer. my inbox is filling up. Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
Ok so i want the library as Free and under the GPL License. I dont want to work with Octave since it is interpreter based and the whole point behind numerical analysis is super speed. If other people want to work on it with me thats fine but i have to say where numerical methods are concerned you have to be good at: 1) math 2) assembly 3) C 4) math Also the university i'm from sukz big time. The Computer Science and Enginnering students dont even know what linux is (go figure). They think that MS visual basic, MS visual .net etc are THE languages. The moral of the story is I have absolutely no faith in other programmers as i have never actaully met other programmers. 6000 students and only 2 people I know are good at programming and they are the only to people who know Linux( anyone see a pattern :)? ) Ok so with the library as Free and License as GPL (or MIT/X or LGPL I dont really care as I've completely given up on understanding the Licence philosophies) where do i goe from here? Somewhere on debian page there was something about uploading and testing. Also if any one can explain to me the licensing issues in 5 3-letter words or less I'll be happy. Hey I might be good at C, assembly, and math*2 but i really suck at legal gibberish. power to the penguin PS: Zen my library is perfect but if you want I'll try my best to add some bugs Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
Actually Zen my goal from the begining was to package it for a GNU/Linux distribution specifically the widely popular and highly recommended Debian. I'll get on making the page at source forge also i'll make a more detailed comparison between octave and my library which i'm calling matlib. i'm comparing it with octave since ocatve seems to be doing the best (most hits for "linux matlab" in google) also thanks for explaining the licences finally i understand. actually i'm not interested in making money from the software. I'm just doing to proove that i can. so i guess i'll go with the GPL. i was just afraid that people would copy-paste the source. (again that happens alot where i'm from) power to the GPL penguin :) Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
On Fri, Feb 06, 2004 at 01:50:50AM -0800, Halim Boukaram wrote:>> I dont want to work with Octave since it is interpreter based and the> whole point behind numerical analysis is super speed.>Old fashion programmer uh? :)Anyway, any C library can be potentially embedded in a scriptinglanguage like octave (as it's true for perl or python). So the integration with octave or scilab or anything else can be attained later.-- Francesco P. LovergineYes ofcourse you are right on both acounts. I am an old fashoined programmer and yes it can be embeded into octave actaully chances are thats how itll end up but i still say interpreters must be avoided for numerical analysis. Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
>If you want it to be widely usable, then I would also suggest packaging >your software in .rpm format. I know a lot of people who work as >sys-admins on Linux systems, and they feel much better when software is in >.rpm, because that's what they know. yes but i'll have to research how to make a good rpm. but i dont get it. i dont like rpms or even makefiles. i stick with good old scripting for compilation stuff. >Yes, and put on your page a very detailed description of what your >software does, what platforms it runs on, and so on. It would also be a >good idea to detail what tests you've run that indicate your software is >better than what exists. since it is C then no doubt it is faster than octave but ofcourse i'll have a detailed description.>I advise you to find a different name for your software. There are a >couple of things out there that have the name "matlib", and it will create >confusion with "matlab", a commercial product. Pick something that will >stand out, like cheetamath. actually this is my second math library. the first was in C++. but instead of just X for ploting curves it used sdl and opengl. i scrapped it for the C library. the C++ library was called mathsim. >Where _are_ you from, just out of curiosity? i'm from Lebanon. The place I study is called NDU. (Its run by money loving preists). just yesterday they were hanging up posters of .net (UNLEASH THE POWER of .Net) oh GOD i almost fainted.>Good luck in what you're doing. Realize that what you're striving for (a >math library that does everything, completely optimized) is the Holy Grail >of scientific computing. Speaking as one who works in that area, when you >do have a product ready to demo, there will be lots of people interested >in what you have. and thats exactly why im doing it. >Good luck! Sincerely,>Craig Steffen power to the penguin Halim BouKaram Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
>If you want it to be widely usable, then I would also suggest packaging >your software in .rpm format. I know a lot of people who work as >sys-admins on Linux systems, and they feel much better when software is in >.rpm, because that's what they know. yes but i'll have to research how to make a good rpm. but i dont get it. i dont like rpms or even makefiles. i stick with good old scripting for compilation stuff. >Yes, and put on your page a very detailed description of what your >software does, what platforms it runs on, and so on. It would also be a >good idea to detail what tests you've run that indicate your software is >better than what exists. since it is C then no doubt it is faster than octave but ofcourse i'll have a detailed description.>I advise you to find a different name for your software. There are a >couple of things out there that have the name "matlib", and it will create >confusion with "matlab", a commercial product. Pick something that will >stand out, like cheetamath. actually this is my second math library. the first was in C++. but instead of just X for ploting curves it used sdl and opengl. i scrapped it for the C library. the C++ library was called mathsim. >Where _are_ you from, just out of curiosity? i'm from Lebanon. The place I study is called NDU. (Its run by money loving preists). just yesterday they were hanging up posters of .net (UNLEASH THE POWER of .Net) oh GOD i almost fainted.>Good luck in what you're doing. Realize that what you're striving for (a >math library that does everything, completely optimized) is the Holy Grail >of scientific computing. Speaking as one who works in that area, when you >do have a product ready to demo, there will be lots of people interested >in what you have. and thats exactly why im doing it. >Good luck! Sincerely,>Craig Steffen power to the penguin Halim BouKaram Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: Packaging problems Re: Packaging .deb vs .rpm
Hi About the C library for math. I'm making the rpms. but the binary rpm isnt working. its turning out much smaller than the source and only keeps the AUTHOR, README, todo and help documents. Ayone knows if thats normal or not. Since i started trying to make the binary rpm i've made one major overhaul of the source. Halim Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
C library faster than octave
Hi Its the 'C library for numerical analysis' guy again. I've tested the differential equation solver for Ocatve (lsode) vs my own with these results: executing an ode solver 1000 times on my P4 1.8: 22 seconds for Ocatve (i think euler method) 8 seconds for my library's euler method 12 seconds for my library's runge kutta order 2 22 seconds for my library's runge kutta order 4 so my estimation is that my library is at least twice as fast as octave which makes it at least as fast as Matlab (which is also twice as fast as Octave). I'll also tell the Octave team maybe they will want to link up. about the packaging: deb, rpm, key, signature? I pretty much gave up. I'll leave it up to someone else. If anyone's interested my projects homepage is: http://numerical.port5.com Sourceforge are taking their time activating my project. __ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25¢ http://photos.yahoo.com/ph/print_splash
Re: C library faster than octave
> Of course, octave does not use the plain Euler > method. Nobody in their > right mind would do that. The octave doc says the > ODE solvers are > "based on reliable ODE solvers written in Fortran", > so they are > probably both more advanced than even plain > Runge-Kutta. You should > test both accuracy and speed. As a test case, I > suggest you try the > initial value problem > > x' = 1+x^2, x(0)=0 > > and solve for x(1), x(1.5), x(1.57). Of course the > exact solution is > x(t)=tan(t), thus the exact values would be 1.5574, > 14.101 and > 1255.8. > > Best regards, > Lukas Your absolutely right. Octave has a much better technique than Euler even better than Runge-Kutta order 4. I'll have to implement something like the predictor-corrector method or one hell of a high order taylor method. Also i'll forget about solving odes of any form and stick to the standard: dx/dt = f(x,t) Thanks. Regards __ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25¢ http://photos.yahoo.com/ph/print_splash
NML vs Octave
Okay Ive decided on the name NML (Numerical Methods Library) Where Odes are concerned. I've implemented 10 different solving schemes. The 2 most promising are seen below vs Octave executed 100 times the equation dy/dx = y*tan(x) [y = sec(x)] which is even more unstable than dy/dx = 1 + y^2. [y = tan(x)]: Real : y(1.57) = 1255.7659 Octave: y(1.57) = 1255.7708 (2.88 sec) Predictor-corrector: y(1.57) = 1211.35(0.32 sec) or y(1.57) = 1233.83(0.64 sec) Runge kutta order 4: y(1.57) = 1255.7659 (exact) (0.59 sec) Runge kutta 4 rocks. __ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover
Re: NML vs Octave
Hi The Brusselator is a system of odes. I'm implementing the algorithm for a system of odes right now using RK4. All I could find out about the Brusselator equation is that is used a lot in chemistry. You seem to know a lot about odes. Does that extend to all of numerical analysis?. If there is a paper you know about that i can read please let me know. Also I found a GNU project similar to mine the GNU Scientific Library (GSL) www.gnu.org/software/gsl/ Its also a C library for numerical analysis. If anyone knows alot about numerical analysis I'd appreciate if you could compare the GSL with my NML numerical.port5.com As far as I can see, setting up the GSL for comutation looks tougher than for the NML. Regards __ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover
What format package for debian test phase?
Hi I've finished making the rpm for my package. Should i convert it to a 'deb' file (using alien) before trying to get it uploaded to debian test folder or should it be rpm or tarred. Regards __ Do you Yahoo!? SBC Yahoo! - Internet access at a great low price. http://promo.yahoo.com/sbc/
C library faster than octave
Hi Its the 'C library for numerical analysis' guy again. I've tested the differential equation solver for Ocatve (lsode) vs my own with these results: executing an ode solver 1000 times on my P4 1.8: 22 seconds for Ocatve (i think euler method) 8 seconds for my library's euler method 12 seconds for my library's runge kutta order 2 22 seconds for my library's runge kutta order 4 so my estimation is that my library is at least twice as fast as octave which makes it at least as fast as Matlab (which is also twice as fast as Octave). I'll also tell the Octave team maybe they will want to link up. about the packaging: deb, rpm, key, signature? I pretty much gave up. I'll leave it up to someone else. If anyone's interested my projects homepage is: http://numerical.port5.com Sourceforge are taking their time activating my project. __ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25¢ http://photos.yahoo.com/ph/print_splash -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: C library faster than octave
> Of course, octave does not use the plain Euler > method. Nobody in their > right mind would do that. The octave doc says the > ODE solvers are > "based on reliable ODE solvers written in Fortran", > so they are > probably both more advanced than even plain > Runge-Kutta. You should > test both accuracy and speed. As a test case, I > suggest you try the > initial value problem > > x' = 1+x^2, x(0)=0 > > and solve for x(1), x(1.5), x(1.57). Of course the > exact solution is > x(t)=tan(t), thus the exact values would be 1.5574, > 14.101 and > 1255.8. > > Best regards, > Lukas Your absolutely right. Octave has a much better technique than Euler even better than Runge-Kutta order 4. I'll have to implement something like the predictor-corrector method or one hell of a high order taylor method. Also i'll forget about solving odes of any form and stick to the standard: dx/dt = f(x,t) Thanks. Regards __ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25¢ http://photos.yahoo.com/ph/print_splash -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
NML vs Octave
Okay Ive decided on the name NML (Numerical Methods Library) Where Odes are concerned. I've implemented 10 different solving schemes. The 2 most promising are seen below vs Octave executed 100 times the equation dy/dx = y*tan(x) [y = sec(x)] which is even more unstable than dy/dx = 1 + y^2. [y = tan(x)]: Real : y(1.57) = 1255.7659 Octave: y(1.57) = 1255.7708 (2.88 sec) Predictor-corrector: y(1.57) = 1211.35(0.32 sec) or y(1.57) = 1233.83(0.64 sec) Runge kutta order 4: y(1.57) = 1255.7659 (exact) (0.59 sec) Runge kutta 4 rocks. __ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: NML vs Octave
Hi The Brusselator is a system of odes. I'm implementing the algorithm for a system of odes right now using RK4. All I could find out about the Brusselator equation is that is used a lot in chemistry. You seem to know a lot about odes. Does that extend to all of numerical analysis?. If there is a paper you know about that i can read please let me know. Also I found a GNU project similar to mine the GNU Scientific Library (GSL) www.gnu.org/software/gsl/ Its also a C library for numerical analysis. If anyone knows alot about numerical analysis I'd appreciate if you could compare the GSL with my NML numerical.port5.com As far as I can see, setting up the GSL for comutation looks tougher than for the NML. Regards __ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
What format package for debian test phase?
Hi I've finished making the rpm for my package. Should i convert it to a 'deb' file (using alien) before trying to get it uploaded to debian test folder or should it be rpm or tarred. Regards __ Do you Yahoo!? SBC Yahoo! - Internet access at a great low price. http://promo.yahoo.com/sbc/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
C library for numerical analysis and math
This is my second email. will it reach? I've written a fast C library for numerical analysis. -integration -differential equations -root finding -interpolation -fourier analysis -ploting curves -misc functions such as rect, tri etc I need to know where to go from here please answer. my inbox is filling up. Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
Ok so i want the library as Free and under the GPL License. I dont want to work with Octave since it is interpreter based and the whole point behind numerical analysis is super speed. If other people want to work on it with me thats fine but i have to say where numerical methods are concerned you have to be good at: 1) math 2) assembly 3) C 4) math Also the university i'm from sukz big time. The Computer Science and Enginnering students dont even know what linux is (go figure). They think that MS visual basic, MS visual .net etc are THE languages. The moral of the story is I have absolutely no faith in other programmers as i have never actaully met other programmers. 6000 students and only 2 people I know are good at programming and they are the only to people who know Linux( anyone see a pattern :)? ) Ok so with the library as Free and License as GPL (or MIT/X or LGPL I dont really care as I've completely given up on understanding the Licence philosophies) where do i goe from here? Somewhere on debian page there was something about uploading and testing. Also if any one can explain to me the licensing issues in 5 3-letter words or less I'll be happy. Hey I might be good at C, assembly, and math*2 but i really suck at legal gibberish. power to the penguin PS: Zen my library is perfect but if you want I'll try my best to add some bugs Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
Actually Zen my goal from the begining was to package it for a GNU/Linux distribution specifically the widely popular and highly recommended Debian. I'll get on making the page at source forge also i'll make a more detailed comparison between octave and my library which i'm calling matlib. i'm comparing it with octave since ocatve seems to be doing the best (most hits for "linux matlab" in google) also thanks for explaining the licences finally i understand. actually i'm not interested in making money from the software. I'm just doing to proove that i can. so i guess i'll go with the GPL. i was just afraid that people would copy-paste the source. (again that happens alot where i'm from) power to the GPL penguin :) Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
On Fri, Feb 06, 2004 at 01:50:50AM -0800, Halim Boukaram wrote:>> I dont want to work with Octave since it is interpreter based and the> whole point behind numerical analysis is super speed.>Old fashion programmer uh? :)Anyway, any C library can be potentially embedded in a scriptinglanguage like octave (as it's true for perl or python). So the integration with octave or scilab or anything else can be attained later.-- Francesco P. LovergineYes ofcourse you are right on both acounts. I am an old fashoined programmer and yes it can be embeded into octave actaully chances are thats how itll end up but i still say interpreters must be avoided for numerical analysis. Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
>If you want it to be widely usable, then I would also suggest packaging >your software in .rpm format. I know a lot of people who work as >sys-admins on Linux systems, and they feel much better when software is in >.rpm, because that's what they know. yes but i'll have to research how to make a good rpm. but i dont get it. i dont like rpms or even makefiles. i stick with good old scripting for compilation stuff. >Yes, and put on your page a very detailed description of what your >software does, what platforms it runs on, and so on. It would also be a >good idea to detail what tests you've run that indicate your software is >better than what exists. since it is C then no doubt it is faster than octave but ofcourse i'll have a detailed description.>I advise you to find a different name for your software. There are a >couple of things out there that have the name "matlib", and it will create >confusion with "matlab", a commercial product. Pick something that will >stand out, like cheetamath. actually this is my second math library. the first was in C++. but instead of just X for ploting curves it used sdl and opengl. i scrapped it for the C library. the C++ library was called mathsim. >Where _are_ you from, just out of curiosity? i'm from Lebanon. The place I study is called NDU. (Its run by money loving preists). just yesterday they were hanging up posters of .net (UNLEASH THE POWER of .Net) oh GOD i almost fainted.>Good luck in what you're doing. Realize that what you're striving for (a >math library that does everything, completely optimized) is the Holy Grail >of scientific computing. Speaking as one who works in that area, when you >do have a product ready to demo, there will be lots of people interested >in what you have. and thats exactly why im doing it. >Good luck! Sincerely,>Craig Steffen power to the penguin Halim BouKaram Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: C library for numerical analysis and math
>If you want it to be widely usable, then I would also suggest packaging >your software in .rpm format. I know a lot of people who work as >sys-admins on Linux systems, and they feel much better when software is in >.rpm, because that's what they know. yes but i'll have to research how to make a good rpm. but i dont get it. i dont like rpms or even makefiles. i stick with good old scripting for compilation stuff. >Yes, and put on your page a very detailed description of what your >software does, what platforms it runs on, and so on. It would also be a >good idea to detail what tests you've run that indicate your software is >better than what exists. since it is C then no doubt it is faster than octave but ofcourse i'll have a detailed description.>I advise you to find a different name for your software. There are a >couple of things out there that have the name "matlib", and it will create >confusion with "matlab", a commercial product. Pick something that will >stand out, like cheetamath. actually this is my second math library. the first was in C++. but instead of just X for ploting curves it used sdl and opengl. i scrapped it for the C library. the C++ library was called mathsim. >Where _are_ you from, just out of curiosity? i'm from Lebanon. The place I study is called NDU. (Its run by money loving preists). just yesterday they were hanging up posters of .net (UNLEASH THE POWER of .Net) oh GOD i almost fainted.>Good luck in what you're doing. Realize that what you're striving for (a >math library that does everything, completely optimized) is the Holy Grail >of scientific computing. Speaking as one who works in that area, when you >do have a product ready to demo, there will be lots of people interested >in what you have. and thats exactly why im doing it. >Good luck! Sincerely,>Craig Steffen power to the penguin Halim BouKaram Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
Re: Packaging problems Re: Packaging .deb vs .rpm
Hi About the C library for math. I'm making the rpms. but the binary rpm isnt working. its turning out much smaller than the source and only keeps the AUTHOR, README, todo and help documents. Ayone knows if thats normal or not. Since i started trying to make the binary rpm i've made one major overhaul of the source. Halim Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online
