[Developers] visit plugin on Cactus webpage not working with Visit 1.9

Ian Hinder ian.hinder at aei.mpg.de
Sat Jul 24 06:17:24 CDT 2010


On 24 Jul 2010, at 13:11, Ian Hinder wrote:

>
> On 24 Jul 2010, at 12:00, Ian Hinder wrote:
>
>>
>> On 24 Jul 2010, at 00:52, Christian Reisswig wrote:
>>
>>> Hi Ian,
>>>
>>> I have just checked the plugin with your data. I get a smooth plot
>>> without any corruptions.
>>> I am looking at the dataset called "phi.xy.h5" that you send me on
>>> 13th of
>>> May. I am also using Visit 1.12.2. What version are you using?
>>
>> I'm using 1.12.0 - I will try with the newer version.  By the way,
>> the file I sent you was called phi.0.xy.h5; is that the one you were
>> trying?  I attach a plot of what I get when I make a pseudocolour
>> plot with range from 0 to 1.
>
>
> Using 1.12.2 didn't help.  I get the same thing.  Did you try a
> pseudocolour plot?  Did you have to slice the data first?  I would
> have thought it should work without, since it is 2D data already.
>
> The steps I take to make the plot are:
>
> 	visit -o phi.0.xy.h5
> 	Plots -> Pseudocolor -> CTGBASE--phi
> 	Draw
>
> I get the image that I sent.  If I then delete the plot and create it
> again by repeating the same steps, I get a different corrupted image.
> I expect there is some memory not being initialised or copied  
> correctly.
>
> I was trying it on Mac OS.  I have now tried it on Linux and it seems
> to work!  The plot is smooth as you reported.  I will check if there
> are other significant differences between the systems (e.g. HDF5
> versions etc).


Looks like I spoke too soon.  Regenerating the plot on Linux gave  
corruption as on Mac OS.  So it looks like it is a bug which only  
shows up sometimes.  Can you try deleting and recreating the plot and  
see if you also get corruption?

-- 
Ian Hinder
ian.hinder at aei.mpg.de



More information about the Developers mailing list