Thanks Jean and Utkarsh for your input, a long night last night spent converting from multiblock to image algorithm did the trick I had to use UPDATE_EXTENT to get the things come out right.I prefer multiblock to image algorithm on principle that good software doesn't do things behind your back, and my impression is that there is alot of black magic going on nehind the scene to make it hapen for parallel image algorithm, or perhaps that's just my stupidity, lack of documentation or some combination thereof ;) <br>
<br><br><div class="gmail_quote">On Tue, Jun 23, 2009 at 12:49 PM, Jean Favre <span dir="ltr"><<a href="mailto:jfavre@cscs.ch">jfavre@cscs.ch</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="im">>>On Mon, Jun 22, 2009 at 4:51 PM, Bam Ting <<a href="mailto:bampingting@gmail.com">bampingting@gmail.com</a>><br>
>>wrote:<br>
>><br>
>>>I wrote a parallel reader using multiblock dataset. How do I do a<br>
>>>volume<br>
>>>rendering?<br>
<br>
</div>It turns out that I had to program my own filter last week, to<br>
re-assemble multiple grids (each of them vtkUniformGrid), into one big<br>
vtkImageData, to enable volume rendering. What I find interesting is<br>
that you say you wrote your own parallel reader. Instead of building a<br>
multi-block, could you not directly insert the data at the right<br>
sub-extents of a *single* vtkImageData?<br>
<br>
The filter I wrote did nothing more than find the global extents of each<br>
grid, allocating a big vtkImageData, and stuffing the sub-arrays in the<br>
right place. Sounds backwards and against what I suggest you should do<br>
yourself. My problem is that I had to deal with AMR data in memory, with<br>
multiple grids per level. That's also why I am not offering my filter<br>
which relies on the AMR internal structure to find the sub-extents.<br>
Would be useless to you.<br>
<br>
By the way, I might be dumb, but I was never able to use the method<br>
vtkImageData::CopyAndCastFrom() which I guess one should use. I wrote my<br>
own loops over the I,J,K indices to make my filter work.<br>
<br>
just 2 more cents<br>
<font color="#888888"><br>
<br>
--<br>
Jean<br>
Swiss National Supercomputing Centre<br>
<br>
</font></blockquote></div><br>