The ICCG solver in the RIPPLE code
Is anyone familiar with the RIPPLE code from Los Alamos National Laboratory?
I have a question about the ICCG solver in it. In the comment lines, words like "periodic diagonals" and "periodic blocks" are mentioned. What do they mean? Thank you in advance! Zach |
I am NOT familiar with the RIPPLE code, but I have seen a matrix A = [a(i,j)] called a periodic Toeplitz matrix with period p, if a(i+p,j+p) = a(i,j) for all (i,j). In particular, a Toeplitz matrix could be considered periodic with period 1.
My best GUESS is that the diagonals in question are periodic, i.e. d(i) = d(i+p), and that the blocks are periodic Toeplitz matrices in the above sense. A few people will also call a tridiagonal matrix with a non-zero entries in the (1,n) and the (n,1) corner 'periodic', presumably because this is the sparsity pattern you get when discretizing, say, the one dimensional heat equation with periodic boundary conditions. /Carl Christian. |
ICCG method is used to solve the linear equation with a symmetric coefficient matrix,you can find the detail on the book of Peric.
|
Thank you, Carl and Xingyue!
|
Hello Zach, you seem familiar using RIPPLE,
i'm now working on developing my own code related to free surface flow, until now i've only had references such as SOLA-VOF and NASA-VOF,and found that RIPPLE is the extended version of them..i'm interested to study it more. i already have the RIPPLE report, but it seems that its source code is not free.. do you have it? if it's not bothering you, could you share the code..i'll be very very thank you if you do.. Regards Novan |
All times are GMT -4. The time now is 17:05. |