Hello, do you know how to use PARDISO-project in VS and Fortran? Since the MKL could not perform iterative calculation and matrix inverse. I have gotten the .lib, .dll and license files, but my code still not work.
Hi! Thank you for the great video. Could you please give me an advice on how to correctly choose iparm to solve five diagonal matrix? High accuracy is preferble to speed.
There are just too many options, depending on the format you use, the type of matrix (whether it is positive-definite), and so on. Just read the manual and set the parameters according to your needs.
Igor, is it possible that your native Pardiso run was not taking advantage of multithreading and that MKL was? I found MKL threading controls vary greatly whether one is using ICC/OpenMP or gcc and TBB.
It seemed like Pardiso was using all cores, as intended - it just performed slower. However, it is possible that the results would be different for different types of matrices.
Sorry, it got accidentally deleted :( I can't find it anymore. Samples that come with MKL and PARDISO show how to invoke the solvers. Sparse matrices can be generated randomly or obtained from "suitesparse matrix collection".
Hello, do you know how to use PARDISO-project in VS and Fortran? Since the MKL could not perform iterative calculation and matrix inverse. I have gotten the .lib, .dll and license files, but my code still not work.
Sorry, I have not used Fortran. You may try to run a C example, then make another shared library from it.
Hello, Igor! Do you still have the code files? I'm interested in reproduce your results with my own matrices. Many Thanks for your video, btw!
sorry - the code seems to be lost :(
Hi! Thank you for the great video. Could you please give me an advice on how to correctly choose iparm to solve five diagonal matrix? High accuracy is preferble to speed.
There are just too many options, depending on the format you use, the type of matrix (whether it is positive-definite), and so on. Just read the manual and set the parameters according to your needs.
Igor, is it possible that your native Pardiso run was not taking advantage of multithreading and that MKL was? I found MKL threading controls vary greatly whether one is using ICC/OpenMP or gcc and TBB.
It seemed like Pardiso was using all cores, as intended - it just performed slower. However, it is possible that the results would be different for different types of matrices.
nice video. I am new to Pardiso, can you tell me how to download it? thanks
It now comes with Intel oneAPI Math Kernel Library (oneMKL) !
@@IgorGribanov Hi can you tell me how to change the value of iparm to enable OOC mode
Hi there! the link for the code and sample files is not working!
Sorry, it got accidentally deleted :( I can't find it anymore. Samples that come with MKL and PARDISO show how to invoke the solvers. Sparse matrices can be generated randomly or obtained from "suitesparse matrix collection".