http://forum.celestialmatters.org/

celestia.Sci and Gaia Sky data (601 million stars)
http://forum.celestialmatters.org/viewtopic.php?f=11&t=932
Page 1 of 1

Author:  sjohn [ Tue, 08-05-18, 18:41 GMT ]
Post subject:  celestia.Sci and Gaia Sky data (601 million stars)

As you may probably all know by now, the huge (601 million stars) ESO Gaia star database is available online.

https://zah.uni-heidelberg.de/institutes/ari/gaia/outreach/gaiasky/downloads/#dr2catalogs

Gaia Sky software is working flawlessly with almost 90GB star data.

https://zah.uni-heidelberg.de/institutes/ari/gaia/outreach/gaiasky/

Is there any plan to integrate these data into celestia.Sci?

Author:  t00fri [ Tue, 08-05-18, 20:01 GMT ]
Post subject:  Re: celestia.Sci and Gaia Sky data (601 million stars)

sjohn wrote:
As you may probably all know by now, the huge (601 million stars) ESO Gaia star database is available online.

https://zah.uni-heidelberg.de/institutes/ari/gaia/outreach/gaiasky/downloads/#dr2catalogs

Gaia Sky software is working flawlessly with almost 90GB star data.

https://zah.uni-heidelberg.de/institutes/ari/gaia/outreach/gaiasky/

Is there any plan to integrate these data into celestia.Sci?


As you presumably have read, our team member Andrew (aka ajtribick) has returned after a long break in real life. Since he's our expert in star extraction we'll certainly implement a reasonable set of DR2 stars into celestia.Sci. Just stay tuned, we'll surely discuss about this matter also in public.

Fridger

Author:  ajtribick [ Wed, 09-05-18, 17:49 GMT ]
Post subject:  Re: celestia.Sci and Gaia Sky data (601 million stars)

Processing the DR2 will require a lot of work, we can't load the entire star database into memory any more because it would be slow and require too much RAM. As an initial step I'm looking at processing the subset of stars that have TYC2 and HIP identifiers, which is similar to the set of stars in the DR1 TGAS set. I will update the 2 million stars thread when I have some results to show.

Author:  sjohn [ Wed, 09-05-18, 18:13 GMT ]
Post subject:  Re: celestia.Sci and Gaia Sky data (601 million stars)

As I understand Gaia Sky loads only a few hundred thousand stars into the RAM, and after jumping to another location it quickly loads another few thousand local stars into the memory.

It would be fantastic to explore faraway regions of our galaxy, because the Gaia mission was able to map some very interesting "corridors" into the other side of the Milky Way.

Is it possible to use the Gaia Sky data somehow "on the fly"?

Author:  t00fri [ Wed, 09-05-18, 19:42 GMT ]
Post subject:  Re: celestia.Sci and Gaia Sky data (601 million stars)

To process 601 million stars in one chunk is clearly out of the question...The optimal number of accounted stars along with their subdivisions needs to be discussed from various aspects.

Here are two of my first thoughts (apart from improving the 2million DR1 star distro):

A) Exploit an Octree structure for processing, just as Celestia/celestia.Sci
already uses it for very fast star display...If the Celestia code appears too complicated,
these GPL Octree C++ sources are quite instructive:
http://nomis80.org/code/octree.html. Have a look at what aggregates are.

B) As a potentially useful analogue consider my Texture Tools that generate 2D tiles very
fast and notably with clever memory management. Hundreds of MB of initial texture
data are no problem whatsoever. The resulting VT tiles are quite analogous to the 3D
Octrees.

Cheers,
Fridger

Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group
http://www.phpbb.com/