[Paraview] XDMF help (solved)
Christine Corbett Moran
corbett at physik.uzh.ch
Mon Aug 24 12:06:12 EDT 2009
Hi Ken,
Thanks! That worked, I had only tried structured topologies which was my
mistake.
Cheers,
Christine
On Mon, Aug 24, 2009 at 5:57 PM, Leiter, Kenneth (Cont, ARL/CISD) <
kenneth.leiter at arl.army.mil> wrote:
> Classification: UNCLASSIFIED
> Caveats: NONE
>
> Hi Christine,
>
> If you are reading in a position array and are unsure of the topology type,
> you will probably want to read in the data with Polyvertex topology, which
> is a group of unconnected points.
>
> The Topology element must specify the number of elements to read. In this
> case, it would be the number of points = 3751383.
>
> This is what I imagine your Xdmf file should look like:
>
> <?xml version="1.0" ?>
> <!DOCTYPE Xdmf SYSTEM "Xdmf.dtd" []>
> <Xdmf xmlns:xi="http://www.w3.org/2003/XInclude" Version="2.2">
> <Domain>
> <Grid>
> <Topology TopologyType="Polyvertex" Dimensions="3751383"/>
> <Geometry>
> <DataItem Dimensions="3751383 3" NumberType="Float" Precision="8"
> Format="HDF">examplefile.h5:/dark/position</DataItem>
> </Geometry>
> </Grid>
> </Domain>
> </Xdmf>
>
>
> - Ken
>
> ---------------------------------------------------------
> Kenneth Leiter
> Scientific Visualization Team
> Lockheed Martin / ARL DoD Supercomputing Resource Center
> 410-278-2580
>
> -----Original Message-----
> From: paraview-bounces at paraview.org [mailto:paraview-bounces at paraview.org]
> On Behalf Of Christine Corbett Moran
> Sent: Monday, August 24, 2009 11:27 AM
> To: paraview at paraview.org
> Subject: [Paraview] XDMF help
>
> Hi,
>
> I have an existing HDF5 file (which I did not personally create so am not
> familiar with other than the info I get from h5dump/h5ls) I'd like to read
> in with Paraview. I'm first trying to create the most basic xmf file
> possible which allows Paraview to read in the position array from the HDF5
> file (based on the documentation here:
> http://www.xdmf.org/index.php/XDMF_Model_and_Format). Right now whatever I
> try either causes a segfault or reads in data nonsensically. Could anyone
> point me in the right direction--e.g. based on the info below are my data
> dimensions just off?--or to better documentation? Thanks in advance, it's
> hopefully something simple I just missed from the documentation.
>
> More information about the file and what I have tried is below.
>
> *After reading a bit and playing around with the parameters, my current
> .xmf
> attempt segfaults shortly after the server prints "Done Parsing." If I
> change TopologyType to 3DSMesh it doesn't segfault but what it manages to
> read in (only a single point) is nonsensical. Any of the other
> TopologyTypes
> listed in the documentation I have found likewise produce a segfault. I
> also
> tried various permutations of the Dimensions argument ("11254149 3", "3
> 11254149", "3751383 3", "3751383 3"), with each possible 3D Topology, these
> likewise manage to read in either just a single point or segfault. I didn't
> try modifying the precision argument as hd5ls tells me the position array
> should be of doubles. This is my current attempt:
> <?xml version="1.0" ?>
> <!DOCTYPE Xdmf SYSTEM "Xdmf.dtd" []>
> <Xdmf xmlns:xi="http://www.w3.org/2003/XInclude" Version="2.2">
> <Domain>
> <Grid>
> <Topology TopologyType="3DRectMesh"/>
> <Geometry>
> <DataItem Dimensions="11254149 3" Precision="8"
> Format="HDF">examplefile.h5:/dark/position</DataItem>
> </Geometry>
> </Grid>
> </Domain>
> </Xdmf>
>
> *About the portion of the file I am trying to read in:
> % h5ls -rv examplefile.h5
> Opened "examplefile.h5" with sec2 driver.
>
> /dark Group
> Attribute: iOrder {1}
>
> Type: native unsigned long
> Data: 0
>
> Location: 0:1:0:2552
> Links: 1
>
> ...
> /dark/position Dataset {11254149/Inf, 3/3}
>
> Location: 0:1:0:2600
> Links: 1
>
> Modified: 2008-07-31 04:12:06 CEST
> Chunks: {32768, 1} 262144 bytes
>
> Storage: 270099576 logical bytes, 270536736 allocated bytes, 99.84%
> utilization
> Filter-0: fletcher32-3 {}
>
> Type: native double
> ...
>
> *More information about the portion of the file I am trying to read in
> obtained from h5dump
>
> %h5dump -x examplefile.h5 > examplefile.xml %grep -A 30 position
> examplefile.xml
>
> <hdf5:Dataset Name="position" OBJ-XID="xid_2600-0" H5Path=
> "/dark/position" Parents="xid_2552-0" H5ParentPaths="/dark">
>
> <hdf5:StorageLayout>
> <hdf5:ChunkedLayout Ndims="2">
>
> <hdf5:ChunkDimension DimSize="32768" />
> <hdf5:ChunkDimension DimSize="1" />
>
> <hdf5:RequiredFilter>
> <hdf5:Fletcher32 />
> </hdf5:RequiredFilter>
>
> </hdf5:ChunkedLayout>
> </hdf5:StorageLayout>
>
> <hdf5:FillValueInfo FillTime="FillIfSet"
> AllocationTime="Incremental">
> <hdf5:FillValue>
>
> <hdf5:NoFill/>
> </hdf5:FillValue>
>
> </hdf5:FillValueInfo>
> <hdf5:Dataspace>
>
> <hdf5:SimpleDataspace Ndims="2">
> <hdf5:Dimension DimSize="11254149" MaxDimSize="UNLIMITED"/>
>
> <hdf5:Dimension DimSize="3" MaxDimSize="3"/>
> </hdf5:SimpleDataspace>
>
> </hdf5:Dataspace>
> <hdf5:DataType>
>
> <hdf5:AtomicType>
> <hdf5:FloatType ByteOrder="LE" Size="8" SignBitLocation="63"
> ExponentBits="11" ExponentLocation="52" MantissaBits="52"
> MantissaLocation="0" />
>
> </hdf5:AtomicType>
> </hdf5:DataType>
>
> <hdf5:Data>
> <hdf5:DataFromFile>
>
> -0.0117184 -0.00222114 -0.0762903
> -0.0111324 -0.00176135 -0.0765049
>
> -0.0106262 -0.00136187 -0.0765845
> -0.0112606 -0.00112936 -0.0766193
>
> ...
> Classification: UNCLASSIFIED
> Caveats: NONE
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20090824/23a0776f/attachment.htm>
More information about the ParaView
mailing list