Hi! I need to use matrix libraries in C++. I see that there are many in oonumerics.com, but please can anyone pin-point an easy-to-use one, and tell me how to use it? It appears the information on the documentation is not enough. I need a step-by-step instructions.
What functions do you need?
try this for basic starters...
you can also look up open source math library and see if any meets your needs.
You can try the LASpack library
But the problem is it is a C based library, but I have used it from a C++ class by writing a very simple wrapper.
Its quite simple and extremely powerful.Feel free to ask if you have some doubts.
Thanks for your tip. I have downloaded lapack. I am using C++.Net. I have incuded the .h files, and LaGenMatDouble A() works, but LaGenMatDouble A(8,8) does not work. The problem might be that I do not have the ddl's, because I looked in the directories and there is none. How do I generate the necessary DDL's. More precisely, can you summarise the instructions for installing and using Lapack? The instructions on the site are a bt ambiguous to me.
Lapack is a pain to get working off unix. There is something wrong with the code or the instructions for using it. For fairly cheap, you can get the Intel math kernel library which is basically an intel chipset optimized (but works on any compatible) version of lapack, blas, etc. Well worth the money, as its fast, documented, and just a couple of #includes away. Alternately, you can try building it on cygwin and moving it into a dll that you can use with .net.
I am not using unix. It is windowsXP prof
oops there seems to be a confusion here...i said LASpack and I think you understood Lapack. I just checked and there is indeed a matrix mulitplication library called lapack :eek:
Sadly I don't have ne experience with lapack, and I have only used LASpack in Unix. I am not sure about windows :(. If you need any help with compiling a program using laspack I can send you the makefile and stuff
Right, your not using unix -- so buy the compiled one for windows was my suggestion, or try to build the lib under cygwin (unix emulator for win32). I didnt notice the s in the name, my brain just read it as lapack. Lapack (no S) is the one intel re-packaged.
Likely the makefile will not help much for a windows compiler, but you never know until you try it. Again, cygwin can save the day here -- well written makefiles work often enough to give it a try.