Hi,
I have a question about running time of Sparse 1.4b.
I tried to solve Ax=b with A being a 10^4 x 10^4 matrix and the number of nonzero elements ~ 5x10^4, and
it worked successfully,
but it took about 10 minutes on the machine (AMD, CPU 2GHz, RAM 1G).
Is there any way to reduce the running time ?
Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sparse is not optimized for factoring a single matrix of a given structure. When it first factors a matrix it determines the ordering and builds its internal data structures. It is much faster on all subsequent factorizations of matrices with the same structure. Thus it is suitable for use in an ODE solver, where you tend to see thousands of matrices of identical structure.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am interested in solving many similar systems. however I am not sure about the spStripFills() function. It is not called in any sparse routines. I tried to added it in the test routine, and got panics. Anyone have an example of how to use it?
examples on add/delete rows/columns are also welcome!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
spStripFills would be called after the matrix has been factored with either spOrderAndFactor or spFactor. If you need more help, give some detail on the panics. There should be no panics, just assertion errors that should indicate what you are doing wrong.
-Ken
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I have a question about running time of Sparse 1.4b.
I tried to solve Ax=b with A being a 10^4 x 10^4 matrix and the number of nonzero elements ~ 5x10^4, and
it worked successfully,
but it took about 10 minutes on the machine (AMD, CPU 2GHz, RAM 1G).
Is there any way to reduce the running time ?
Thanks in advance.
Sparse is not optimized for factoring a single matrix of a given structure. When it first factors a matrix it determines the ordering and builds its internal data structures. It is much faster on all subsequent factorizations of matrices with the same structure. Thus it is suitable for use in an ODE solver, where you tend to see thousands of matrices of identical structure.
I am interested in solving many similar systems. however I am not sure about the spStripFills() function. It is not called in any sparse routines. I tried to added it in the test routine, and got panics. Anyone have an example of how to use it?
examples on add/delete rows/columns are also welcome!
spStripFills would be called after the matrix has been factored with either spOrderAndFactor or spFactor. If you need more help, give some detail on the panics. There should be no panics, just assertion errors that should indicate what you are doing wrong.
-Ken