Code to reproduce the numerical experiments in "Fast global spectral methods for three-dimensional partial differential equations" by C. Strössner and D. Kressner.
MIT License
5
stars
0
forks
source link
Question about CP decomposition in "OpSplittingNonConst" function in the code "PlotHelmholtzUnknown.m" #1
I have a question regarding the implementation of your paper's algorithm for solving the Helmholtz equation using tensor decomposition. In section 5.2.4, Remarks, it is mentioned that for n1=n2=n3=30, the algorithm does not solve CP decomposition for (31x3) x (31x3) x (31x3), but instead approximates the operators ∆ and κ(x, y, z)I (which have size of 31x31x31) separately. However, I could not find this in the code "PlotHelmholtzUnknown.m". Can you please confirm whether the code implements the approximation mentioned in the paper?
Dear authors,
I have a question regarding the implementation of your paper's algorithm for solving the Helmholtz equation using tensor decomposition. In section 5.2.4, Remarks, it is mentioned that for n1=n2=n3=30, the algorithm does not solve CP decomposition for (31x3) x (31x3) x (31x3), but instead approximates the operators ∆ and κ(x, y, z)I (which have size of 31x31x31) separately. However, I could not find this in the code "PlotHelmholtzUnknown.m". Can you please confirm whether the code implements the approximation mentioned in the paper?
Thank you for your time and assistance.