topic d_commit_Helmholtz_3D in Intel® oneAPI Math Kernel Library

d_commit_Helmholtz_3D

Ahmad_Falahatpisheh — Tue, 29 May 2012 06:10:08 GMT

Dear All,

I have a question on using d_commit_Helmholtz_3D and d_Helmholtz_3D. I would like to know if the right hand side of the Poisson's equation needs an FFT before assigning to "f" argument of these functions.

Thanks,

d_commit_Helmholtz_3D

Alexander_K_Intel2 — Tue, 29 May 2012 07:59:02 GMT

Hi,

I don't understand you question could you reformulate it? Nevertheless some days ago we published KB article about using PL and TT components of MKL (http://software.intel.com/en-us/articles/intel-mkl-poisson-library-and-trigonometric-transform-overview/), maybe it will be useful.

With best regards,

Alexander Kalinkin

d_commit_Helmholtz_3D

Ahmad_Falahatpisheh — Tue, 29 May 2012 17:01:37 GMT

Dear Alexander,

I need to assign the values of "f", the RHS of Poisson's equation, in three nested for-loops before calling Helmholtz 3D functions. The Poisson's equation is U=f(x,y,z). My question is: should I do an FFT on f(x,y,z) and then assign the transformed values of f(x,y,z) to the "f" argument of Helmholtz functions? Or I simply use the values of f(x,y,z) without any FFT for Helmholtz functions?

Thanks.

d_commit_Helmholtz_3D

Alexander_K_Intel2 — Tue, 29 May 2012 18:50:47 GMT

Hi,

The last variant is correct; you just need to provide your rhs to commit and main routines of Helmholtz and it calculate solution of linear equation system arises from grid Helmholtz equation.

With best regards,

Alexander Kalinkin