NEO tracking script beta2

[amaury]

NEO tracking script beta2 is out! (i felt this one deserved another post, i hope nobody gets angry )


http://asolignac.free.fr/celestia/NWSv2beta.celx


Interface has been improved, output is not mandatory anymore (but the html file now includes cel: // urls), and progressive zooming has been added for asteroids. Most important, there is now a help and info screen (thanks to Selden for his questions)

i am curious to know how Celestia runs this script on other machines before i release the script.
With Arlene's NEO and Dirl's comets ssc files, I get something around and 23fps for a comet sim?° with or without html output. For an asteroid sim?°: 6fps with output, 7 without, which is still comfortable.

Next step would be to compute the relative speed and inclination of hypothetical impactors (hence my post on relative speed/velocity).
But i might as well release it as it is.


Thanks for comments and suggestions!
amaury


Dirl's comets
http://www.celestiamotherlode.net/catalog/show_addon_details.php?addon_id=788
Arlene's asteroids
http://www.celestiamotherlode.net/catalog/show_addon_details.php?addon_id=646

[selden]

You seem to have reversed the links to the Comet and Asteroid Addons.

I get about 6-8 fps for Comets, about 3.8 for Asteroids with no output.

System:
512MB 2.4GHz P4, WinXP SP2
128MB GF 5200FX, Forceware v66.81
Celestia v1.4.1 final

[amaury]

gloups, that's what i feared

too many loops probably.
Can i ask you to try the first version for fps
http://asolignac.free.fr/celestia/NWSv1beta.celx
and see if there is any difference ?

My own system is :
1GB 3GHz PD(hyperthread), WinXP SP2
128MB Radeon X600
Celestia v1.4.1 final

thanks

[selden]

My previous values were with both comet and asteroid databases loaded. Most of the slowdown is due to the large number of asteroid orbits that Celestia is calculating.

With only comets loaded, I get frame rates of about 28 fps for both scripts while they search for near-Earth comets.

Sorry for the confusion.

[amaury]

i feel it almost becomes a nuisance if one has to enable/disable some ssc files before using the script. Doesn't really fit with the idea of an "educational", easy script

I guess if i upload it, i should mention it works well for comets and small asteroid files, but does require a fast computer for large asteroid description files... (Arlene's NEO file has over 3000 objects... this explains that)

Thanks for trying out and finding this Selden!

By the way, the OrbitPathSamplePoints in celestia.cfg doesn't seem to have any impact on accuracy or performance of the script (only a visual render thing?)
And, one last question : what is the celestia window size when you try ?
(it has some effect on performance for my configuration)

Any other feedback form anyone is welcome

cheers
amaury

[selden]

Orbit paths only have to be calculated once since they don't move. After that, the work in determining where they're located relative to the viewpoint is done by the grahics card. The locations of orbiting bodies have to be calculated for every frame since they do move.

I was running in a window about 1K on a side. Full screen (1600x1200) reduces the comet speed to about 15fps. Asteroids are still about 3.5fps.

[amaury]

so i might end up adjusting the number of mainloops to the framerate, using a "mainloops per second" indicator that is already in the script (the "mps" variable).

This way, the script would perform fewer mainloops (1 instead of 2 for instance), with reduced accuracy on distances calculation, but better comfort for users.

thanks

Amaury