[vtkusers] Parallel/multi-threaded FLUENT VTK reader

[Tamás Egeresi]

Dear All,

I've written a serial FLUENT post-processing code in Python using its 
VTK module, and just found out that the code reaches the limit of the 
readable file size. My sample FLUENT file consists of 1.5 millions of 
cells, and the serial code runs with it smoothly producing a matplotlib 
surface plot, but with another mesh with 4 millions cells it fails to 
run, giving "Kernel died unexpectedly ...". My computer has 8 CPUs, 16GB 
RAM and runs with Windows 10, so I'm about to use threaded version of 
the reader. My question is if it is adaptable or not, and if yes, how it 
is possible to be implemented.

I have read a vast amount of material focusing on parallel filters, but 
my issue is different since I need the parallelism (or threading) right 
in the reader section, which is the bottleneck of the problem. I've  
also tried to start developing the reader by using pickable vtk 
function, but could not find any appropriate information of its usage. 
Some details of it, or any homepage is very welcome.

The only solution has been up till now that I use ParaVIEW to produce 
slices in separate files, and run my serial FLUENT reader on them. My 
goal is to exclude the ParaVIEW step and to work in Python, only.

-- 
T. Egeresi
Design engineer