[yt-dev] Interacting with data in yt 3.0 (was Field units from code to code)

[Matthew Turk]

Hi Sam,

On Tue, Apr 3, 2012 at 3:05 PM, Sam Skillman <samskillman at gmail.com> wrote:
> Back to the point...
>
> That looks great, Matt.  I'll get some stuff going on libyt/gdf when I get
> the chance.  I will also return to finishing up my part of the volume
> rendering refactor, which I think just needs to be touched up and would
> probably be ready for something more like 2.4.

Awesome!  I think this definitely should aim for 2.4.  For those of
you reading along that haven't been following the vr-refactor
bookmark, the basic idea is that volume rendering is now separated
into traversal and accumulator functions.  The accumulator functions
are all written in nogil Cython, which allows them to be threaded.
Additionally, operations like off-axis projections now do not
interpolate (or evaluate the transfer function) and so they operate
much faster.  So combine both the optimized accumulator/samplers with
threading, and you get much improved speed.  (For off-axis projections
-- again, no longer interpolated -- this translated in my tests to
~20x.)  Also, for some operations (out-of-order, which include the old
style volume renderings with no alpha blending!) no longer does the
volume need to be homogenized; it can just be fed in directly with
vertex centered data.

Sam's also done a lot of work to improve the functionality of the
volume renderer; last I remember he had a pretty nice looking
opaque-surface with specular lighting.

When 2.4 goes out we should also include some benchmarking data that
shows off the speed increase, as well as the new capabilities.

-Matt

>
> I also had some thoughts on the parameter file vs dataset, but perhaps I'll
> move those comments to the doc.
>
> Sam
>
>
> On Tue, Apr 3, 2012 at 12:56 PM, Nathan Goldbaum <goldbaum at ucolick.org>
> wrote:
>>
>> http://philosophistry.com/archives/2009/01/what-is-a-black.html
>>
>> Nathan Goldbaum
>> Graduate Student
>> Astronomy & Astrophysics, UCSC
>> goldbaum at ucolick.org
>> http://www.ucolick.org/~goldbaum
>>
>> On Apr 3, 2012, at 11:54 AM, Stephen Skory wrote:
>>
>> PS I had a "black triangle" moment today with octree reading, where I
>>
>> was able to apply geometric selection using the same mechanisms as
>>
>> grid patches, but without the expensive regridding step.  Hooray!
>>
>>
>> Um, I'm not familiar with what you mean by "black triangle", I think
>> it means you had a successful moment when things all came together,
>> but I'm guessing it's not this?
>>
>> https://en.wikipedia.org/wiki/Black_triangle_(badge)
>>
>>
>> --
>> Stephen Skory
>> s at skory.us
>> http://stephenskory.com/
>> 510.621.3687 (google voice)
>> _______________________________________________
>> yt-dev mailing list
>> yt-dev at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>
>> !DSPAM:10175,4f7b47742854955923471!
>>
>>
>>
>> _______________________________________________
>> yt-dev mailing list
>> yt-dev at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>
>
>
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>