[yt-users] volume rendering with artio

[Matthew Turk]

Hi Owen,

Thanks for the bug report -- I tracked it down and it was a pernicious
difference between the number of cells and the number of octs.  It
crept in while I was redoing the ARTIO frontend with Doug a few weeks
ago.  (I was the one who made the mistake, not Doug.)

I've pushed a fix to:

http://bitbucket.org/yt_analysis/yt-3.0

in changeset c86499f6891e.  I was able to off-axis project a sample
dataset, although volume rendering failed with what look like indexing
errors because ARTIO has (1,1,1) patch datasets at the root mesh
level, which are confusing the kD-tree.  Testing against regressions
in this type of thing is not terribly hard, so I am adding that to the
testing suite and will include it in the updates to that that I'm
working on this afternoon.  I should note that as it stands, the
volume rendering of even very small ARTIO data can be very, very slow,
as the VR algorithm is optimized for larger patches.  That will likely
not be fixed for a while, unfortunately, although addressing volume
rendering for non-patch datasets is a priority.

-Matt

On Fri, Oct 4, 2013 at 3:08 PM, Owen Parry <oparry at umd.edu> wrote:
> No problem, thanks for your rapid response!
>
> Owen
>
>
> On 04/10/2013 15:05, Matthew Turk wrote:
>>
>> Hi Owen,
>>
>> I'm going to look into this, but I think for now probably it's safe to
>> say VR is not working for octree data, and it will be high on my list
>> of things to fix.  I have verified the issue and will take a look as
>> soon as I am able.
>>
>> Sorry for the trouble.
>>
>> -Matt
>>
>> On Fri, Oct 4, 2013 at 3:00 PM, Owen Parry <oparry at umd.edu> wrote:
>>>
>>> Hi Matt,
>>>
>>> The version hash is c1540c8d27cb.
>>> I think I was using no_ghost = True as the default anyway, but I get the
>>> same result specifying it explicitly.
>>>
>>> Owen
>>>
>>>
>>>
>>> On 04/10/2013 14:51, Matthew Turk wrote:
>>>>
>>>> Hi Owen,
>>>>
>>>> What's the version hash output by "yt instinfo"?  You may also be
>>>> restricted at the present time to things that don't require ghost
>>>> zones, which would be the "no_ghost = True" option to the Camera.
>>>>
>>>> -Matt
>>>>
>>>> On Fri, Oct 4, 2013 at 2:49 PM, Owen Parry <oparry at umd.edu> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> I'm trying to do some volume rendering of artio outputs, but am running
>>>>> into
>>>>> problems setting up the camera.
>>>>> Seems like pf.h.grids is used when building the kd tree, but isn't
>>>>> defined
>>>>> for artio...?
>>>>> I'm using yt-3.0.
>>>>>
>>>>> Traceback here:
>>>>> http://paste.yt-project.org/show/2ouFxUS2MrGXbhxjLD6L/
>>>>>
>>>>> My script is pretty much identical to the one in the volume rendering
>>>>> tutorial:
>>>>> http://paste.yt-project.org/show/3921/
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Owen
>>>>> _______________________________________________
>>>>> yt-users mailing list
>>>>> yt-users at lists.spacepope.org
>>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>>
>>>> _______________________________________________
>>>> yt-users mailing list
>>>> yt-users at lists.spacepope.org
>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>
>>>
>>> _______________________________________________
>>> yt-users mailing list
>>> yt-users at lists.spacepope.org
>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>
>> _______________________________________________
>> yt-users mailing list
>> yt-users at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org