5.6 FieldConvert in parallel

To run FieldConvert in parallel the user needs to compile Nektar++ with MPI support and can employ the following command

mpirun -np <nprocs> FieldConvert test.xml test.fld test.dat
mpirun -np <nprocs> FieldConvert test.xml test.fld test.plt

or

mpirun -np <nprocs> FieldConvert test.xml test.fld test.vtu

replacing <nprocs> with the number of processors. For the .dat and .plt outputs the current version will proudce a single output file. However it is also sometimes useful to produce multiple output files, one for each partition, and this can be done by using the writemultiplefiles option, i.e.

mpirun -np <nprocs> FieldConvert test.xml test.fld \ 
                test.dat:dat:writemultiplefiles
mpirun -np <nprocs> FieldConvert test.xml test.fld \ 
                test.plt:plt:writemultiplefiles

For the .vtu format multiple files will by default be produced of the form test_vtu/P0000000.vtu, test_vtu/P0000001.vtu, test_vtu/P0000002.vtu. For this format an additional file called test.pvtu is written out which allows for parallel reading of the individual .vtu files.

FieldConvert functions that produce a .fld file output will also be created when running in parallel. In this case when producing a .fld file a directory called test.fld (or the specified output name) is created with the standard parallel field files placed within the directory.