Hi Cheng,

I hope to be corrected on what I write here, but I think parallel decomposition in the Fourier direction works only in the 3DH1D case. I was using the 1D mesh with 2D Fourier expansion in parallel, but decomposition needed to be done on the 1D mesh. As soon as you try to use --npz the simulation will not run. Also note that in the 3DH2D you will not to be able to use some of the, otherwise available filters. I ended up using a 2D mesh with 1D Fourier.

Cheers,

Stan

8 maj 2021 22:39 "Xiao, Chengnian" <CHX33@pitt.edu> napisał(a):

This email from CHX33@pitt.edu originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list to disable email stamping for this address.

Dear all,

I would like to know if the latest nektar++ version supports parallel 3DH2D simulation,

i.e. 3D configuration based on 1D mesh and Fourier expansions in y,z directions?

According to the user-guide, there are ---npz, -npy,--npx options available. However,

whenever I run the 3DH2D test case provided in the sample session file

ChanFlow_3DH2D_FFT.xml on multiple processors with mpirun -np >1, the solver

would crash after initialization, no matter if npx, npy or npz are specified to be greater

than 1 or not. The case does work fine however on a single processor.

Best regards,

Cheng